Last active
September 15, 2020 20:30
-
-
Save ZihengJiang/6d3440ec22852dc9baae2e3f278ad8b4 to your computer and use it in GitHub Desktop.
[WIP] Apache TVM (incubating) v0.7 Release Notes
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| ## Dec 2019 | |
| ### Compiler Support | |
| * Add function attributes to IR hash ([#4479](https://github.com/apache/incubator-tvm/pull/4479/files)) | |
| * Intrinsic dispatching with OCML instead of LLVM for ROCm ([#4499](https://github.com/apache/incubator-tvm/pull/4499)) | |
| * IR readability enhancement ([#4501](https://github.com/apache/incubator-tvm/pull/4501)) | |
| * Add bfloat16 typeflag support ([#4525](https://github.com/apache/incubator-tvm/pull/4525)) | |
| * External codegen support in Relay ([#4482](https://github.com/apache/incubator-tvm/pull/4482)) + VM ([#4544](https://github.com/apache/incubator-tvm/pull/4544)) | |
| * Deprecating NNVM ([#4535](https://github.com/apache/incubator-tvm/pull/4535), [#4562](https://github.com/apache/incubator-tvm/pull/4562), [#4565](https://github.com/apache/incubator-tvm/pull/4565), [#4571](https://github.com/apache/incubator-tvm/pull/4571)) | |
| * Cythonize NDArray.copyto ([#4549](https://github.com/apache/incubator-tvm/pull/4549)) | |
| * Add convertlayout pass in Relay ([#4335](https://github.com/apache/incubator-tvm/pull/4335), [#4600](https://github.com/apache/incubator-tvm/pull/4600)) | |
| * Relay passes lookup overhead optimization ([#4594](https://github.com/apache/incubator-tvm/pull/4594)) | |
| * Unified Object System runtime refactor ([#4578](https://github.com/apache/incubator-tvm/pull/4578), [#4581](https://github.com/apache/incubator-tvm/pull/4581), [#4603](https://github.com/apache/incubator-tvm/pull/4603)) | |
| * VM profiler: sort VM stats by time ([#4601](https://github.com/apache/incubator-tvm/pull/4601)) | |
| ### Operator Support and AutoTVM | |
| * Add strided_set operation ([#4303](https://github.com/apache/incubator-tvm/pull/4303)) | |
| * Add shape function for zero, zeros_like, ones, ones_like ([#4448](https://github.com/apache/incubator-tvm/pull/4448)), tile ([#4441](https://github.com/apache/incubator-tvm/pull/4441/files)) | |
| * Add support for conv3d ([#4400](https://github.com/apache/incubator-tvm/pull/4400)), pool3d ([#4478](https://github.com/apache/incubator-tvm/pull/4478)), 3d upsampling ops ([#4584](https://github.com/apache/incubator-tvm/pull/4584)) | |
| * Add group convolution for VTA ([#4421](https://github.com/apache/incubator-tvm/pull/4421)) | |
| * Adding ROCM schedules for TOPI ([#4507](https://github.com/apache/incubator-tvm/pull/4507)) | |
| * Add 1d deconvolution op ([#4476](https://github.com/apache/incubator-tvm/pull/4476)) | |
| * Allow batch matmul to be fused into injective ops ([#4537](https://github.com/apache/incubator-tvm/pull/4537)) | |
| * Add native depthtospace and spacetodepth operators ([#4566](https://github.com/apache/incubator-tvm/pull/4566)) | |
| * NHWC conv2d schedule templates for ARM ([#3859](https://github.com/apache/incubator-tvm/pull/3859)) | |
| * Int8 GEMM performance enhancement using Cublas ([#4550](https://github.com/apache/incubator-tvm/pull/4550)) | |
| ### User Interface and Frontend | |
| * TFLite parser support for transpose_conv ([#4440](https://github.com/apache/incubator-tvm/pull/4440)), unpack ([#4447](https://github.com/apache/incubator-tvm/pull/4447)) | |
| * LLDB pretty printers for relay ([#4453](https://github.com/apache/incubator-tvm/pull/4453)) | |
| * ONNX to Relay converter op support: expand op ([#4483](https://github.com/apache/incubator-tvm/pull/4483)), auto_pad in conv and convtranspose ([#4563](https://github.com/apache/incubator-tvm/pull/4563/)) | |
| * TF to Relay converter op support: bilinear and neighbour implementation refactor ([#4504](https://github.com/apache/incubator-tvm/pull/4504)), max_pool3d ([#4551](https://github.com/apache/incubator-tvm/pull/4551)), conv2d_transpose with “same” padding support for larger than 1x1 kernels | |
| * Remove unnecessary cast of constants in ONNX converter ([#4573](https://github.com/apache/incubator-tvm/pull/4573)) | |
| ### Runtime | |
| * Add ADTObject POD container type ([#4346](https://github.com/apache/incubator-tvm/pull/4346)) | |
| * Add CUDNN conv3d support ([#4418](https://github.com/apache/incubator-tvm/pull/4418/files)) | |
| * Update RPC runtime to allow remote module as arg ([#4462](https://github.com/apache/incubator-tvm/pull/4418/files)) | |
| * TFLite RPC runtime ([#4439](https://github.com/apache/incubator-tvm/pull/4439)) | |
| * Refactorying system lib and dso lib into library module ([#4481](https://github.com/apache/incubator-tvm/pull/4481)) | |
| * Standardized graph runtime export ([#4532](https://github.com/apache/incubator-tvm/pull/4532)) | |
| ### Documents, Test, and Build | |
| * Adding benchmark log format doc ([#4366](https://github.com/apache/incubator-tvm/pull/4366)) | |
| * Adding AMD codegen unit tests ([#4509](https://github.com/apache/incubator-tvm/pull/4509)) | |
| * Add Ninja build system to installation docs ([#4554](https://github.com/apache/incubator-tvm/pull/4554)) | |
| * Add v0.6 release ([#4558](https://github.com/apache/incubator-tvm/pull/4558)) | |
| ### Accelerator and Microcontroller Support | |
| * uTVM support for ARM STM32F746XX boards ([#4274](https://github.com/apache/incubator-tvm/pull/4274)) | |
| * Speedup TSIM with multi-threading ([#4491](https://github.com/apache/incubator-tvm/pull/4491)) | |
| * Improve TSIM virtual memory mapping ([#4545](https://github.com/apache/incubator-tvm/pull/4545)) | |
| * Cleanup legacy verilog code ([#4576](https://github.com/apache/incubator-tvm/pull/4576)) | |
| ### Fixes | |
| * Doc/comment fixes ([#4452](https://github.com/apache/incubator-tvm/pull/4452), [#4463](https://github.com/apache/incubator-tvm/pull/4463), [#4469](https://github.com/apache/incubator-tvm/pull/4469), [#4493](https://github.com/apache/incubator-tvm/pull/4493), [#4397](https://github.com/apache/incubator-tvm/pull/4397), [#4580](https://github.com/apache/incubator-tvm/pull/4580), [#4585](https://github.com/apache/incubator-tvm/pull/4585), [#4591](https://github.com/apache/incubator-tvm/pull/4591)) | |
| * MSVC / Windows fixes ([#4455](https://github.com/apache/incubator-tvm/pull/4455), [#4569](https://github.com/apache/incubator-tvm/pull/4569)) | |
| * Fix Makefile for howto_deploy ([#4457](https://github.com/apache/incubator-tvm/pull/4457/files)) | |
| * Fix GCC 4.8 compact ([#4461](https://github.com/apache/incubator-tvm/pull/4461)) | |
| * Fix search path to build libtvm_topi.so ([#4467](https://github.com/apache/incubator-tvm/pull/4467)) | |
| * Fix for conv2d_transpose CUDA compilation ([#4472](https://github.com/apache/incubator-tvm/pull/4472)) | |
| * Fix for LLVM 10.0 codegen ([#4480](https://github.com/apache/incubator-tvm/pull/4480), [#4515](https://github.com/apache/incubator-tvm/pull/4515)) | |
| * Fix alter op layout when calling global var ([#4454](https://github.com/apache/incubator-tvm/pull/4454)) | |
| * Fix float2half_rn support for cuda compute capabilities < 53 ([#4489](https://github.com/apache/incubator-tvm/pull/4489)) | |
| * Fix compile errors for OpenCL backends ([#4492](https://github.com/apache/incubator-tvm/pull/4492)) | |
| * Fix serialization precision loss ([#4503](https://github.com/apache/incubator-tvm/pull/4503)) | |
| * Fix hybrid script to support array of tensors ([#4494](https://github.com/apache/incubator-tvm/pull/4494)) | |
| * Fix annotation for multiply op ([#4458](https://github.com/apache/incubator-tvm/pull/4458)) | |
| * Fix Dockerfile for linter CI ([#4506](https://github.com/apache/incubator-tvm/pull/4506)) | |
| * Fix TF resize for dynamic size models ([#4510](https://github.com/apache/incubator-tvm/pull/4510)) | |
| * Fix bias_add gradient ([#4516](https://github.com/apache/incubator-tvm/pull/4516)) | |
| * Fix tanH unit test function call ([#4517](https://github.com/apache/incubator-tvm/pull/4517)) | |
| * Fix extra reshape parameter for ONNX ([#4524](https://github.com/apache/incubator-tvm/pull/4524)) | |
| * Fix crash caused by empty TOPI config ([#4520](https://github.com/apache/incubator-tvm/pull/4520)) | |
| * Fix ONNX shape op type to use int64 ([#4528](https://github.com/apache/incubator-tvm/pull/4528)) | |
| * Fix crash in TSIM virtual memory driver ([#4527](https://github.com/apache/incubator-tvm/pull/4527)) | |
| * Replace deprecated python library in setup script ([#4533](https://github.com/apache/incubator-tvm/pull/4533)) | |
| * Fix NMS max_output_size loop ([#4541](https://github.com/apache/incubator-tvm/pull/4541)) | |
| * Fix style in IR mutator and IR visitor ([#4561](https://github.com/apache/incubator-tvm/pull/4561)) | |
| * Fix compiler warning ([#4559](https://github.com/apache/incubator-tvm/pull/4559)) | |
| * Fix to get end to end inference on Chisel VTA ([#4574](https://github.com/apache/incubator-tvm/pull/4574)) | |
| * Fix LLVM build by adding missing intrinsics headers ([#4575](https://github.com/apache/incubator-tvm/pull/4575)) | |
| * Fix context creation in quantization ([#4582](https://github.com/apache/incubator-tvm/pull/4582)) | |
| * Fix NDArray SaveDLTensor signature ([#4586](https://github.com/apache/incubator-tvm/pull/4586)) | |
| * Fix dense pack schedule for x86 ([#4539](https://github.com/apache/incubator-tvm/pull/4539)) | |
| * Fix for broadcast tensor of scalar type ([#4577](https://github.com/apache/incubator-tvm/pull/4577)) | |
| * Datatype refactor ([#4513](https://github.com/apache/incubator-tvm/pull/4513/files), [#4560](https://github.com/apache/incubator-tvm/pull/4560)) | |
| * Add const qualifiers for NDArray container ([#4590](https://github.com/apache/incubator-tvm/pull/4590)) | |
| * Fix TF <= 1.12 compatibility ([#4593](https://github.com/apache/incubator-tvm/pull/4593)) | |
| * Fix for graph debug runtime ([#4598](https://github.com/apache/incubator-tvm/pull/4598/files)) | |
| * Disable copy constructor for external codegen ([#4597](https://github.com/apache/incubator-tvm/pull/4597)) | |
| * Make ADT tag signed ([#4605](https://github.com/apache/incubator-tvm/pull/4605)) | |
| ## Jan 2020 | |
| ### Refactor | |
| - Unify approach to Visitor/Mutator under Functor [#4606](https://github.com/apache/incubator-tvm/issues/4606) | |
| - TVM_REGISTER_API -> TVM_REGISTER_GLOBAL [#4621](https://github.com/apache/incubator-tvm/issues/4621) | |
| - Introduce SeqStmt to replace ir::Block [#4627](https://github.com/apache/incubator-tvm/issues/4627) | |
| - IRPrinter->NodePrinter, move to node/printer.h [#4622](https://github.com/apache/incubator-tvm/issues/4622) | |
| - Initialize Unified IR Type Data Structures [#4616](https://github.com/apache/incubator-tvm/issues/4616) | |
| - Remove un-necessary var sub-field in TypeVars [#4615](https://github.com/apache/incubator-tvm/issues/4615) | |
| - Remove old Low-level Visitor/Mutator [#4612](https://github.com/apache/incubator-tvm/issues/4612) | |
| - Migrate Low-level IR Passes into the New Stmt/Expr Mutator [#4607](https://github.com/apache/incubator-tvm/issues/4607) | |
| - Automatically deduce ftype signature in Registry.set_body_typed [#4623](https://github.com/apache/incubator-tvm/issues/4623) | |
| - Add Node suffix to low-level IR nodes [#4649](https://github.com/apache/incubator-tvm/issues/4649) | |
| - relay::Module Def -> TypeDef [#4665](https://github.com/apache/incubator-tvm/issues/4665) | |
| - tvm::Expr -> PrimExpr(Primitive Expr) [#4669](https://github.com/apache/incubator-tvm/issues/4669) | |
| - Replace TensorObj and TensorValue with NDArray [#4643](https://github.com/apache/incubator-tvm/issues/4643) | |
| - Initialize Unified IR Expr Data Structure [#4673](https://github.com/apache/incubator-tvm/issues/4673) | |
| - Allow Module to store BaseFunc [#4678](https://github.com/apache/incubator-tvm/issues/4678) | |
| - Unified IR Primitive Op and Registry [#4687](https://github.com/apache/incubator-tvm/issues/4687) | |
| - Unified IR IRModule structure. [#4699](https://github.com/apache/incubator-tvm/issues/4699) | |
| - Move error.h into ir [#4701](https://github.com/apache/incubator-tvm/issues/4701) | |
| - Initialize Unified IR Pass Infra [#4702](https://github.com/apache/incubator-tvm/issues/4702) | |
| - Polish ir/type [#4705](https://github.com/apache/incubator-tvm/issues/4705) | |
| - Unify IntImm and UIntImm [#4706](https://github.com/apache/incubator-tvm/issues/4706) | |
| - attrs.h -> ir [#4709](https://github.com/apache/incubator-tvm/issues/4709) | |
| - Move support related code to include/tvm/support [#4716](https://github.com/apache/incubator-tvm/issues/4716) | |
| - Make more clear naming for C API Type codes. [#4715](https://github.com/apache/incubator-tvm/issues/4715) | |
| - Introduce include/tvm/target [#4721](https://github.com/apache/incubator-tvm/issues/4721) | |
| - Unified IR, introduce include/tvm/arith/ [#4722](https://github.com/apache/incubator-tvm/issues/4722) | |
| - top - namespace for Tensor Operation DSL [#4727](https://github.com/apache/incubator-tvm/issues/4727) | |
| - Polish runtime [#4729](https://github.com/apache/incubator-tvm/issues/4729) | |
| - Get rid of packed_func_ext. [#4735](https://github.com/apache/incubator-tvm/issues/4735) | |
| - Unify vm and interpreter objects [#4693](https://github.com/apache/incubator-tvm/issues/4693) | |
| - Establish tir namespace [#4740](https://github.com/apache/incubator-tvm/issues/4740) | |
| - codegen->target, build_module->driver [#4742](https://github.com/apache/incubator-tvm/issues/4742) | |
| - Finish move all types to IR. [#4746](https://github.com/apache/incubator-tvm/issues/4746) | |
| - Establish printer in the source folder [#4752](https://github.com/apache/incubator-tvm/issues/4752) | |
| - top->te [#4759](https://github.com/apache/incubator-tvm/issues/4759) | |
| - driver.h -> driver_api.h [#4760](https://github.com/apache/incubator-tvm/issues/4760) | |
| ### Relay | |
| - Add half_pixel option to Resize op [#4610](https://github.com/apache/incubator-tvm/issues/4610) | |
| - skip example json runtime test when config is not set [#4614](https://github.com/apache/incubator-tvm/issues/4614) | |
| - test tensor_array in vm [#4608](https://github.com/apache/incubator-tvm/issues/4608) | |
| - Improve memory_allocation pass to support multiple i/o dynamic kernels [#4595](https://github.com/apache/incubator-tvm/issues/4595) | |
| - add unit test for tensor_array_split [#4619](https://github.com/apache/incubator-tvm/issues/4619) | |
| - Add parses support for unary elemwise ops [#4634](https://github.com/apache/incubator-tvm/issues/4634) | |
| - Add parses support for SLICE [#4502](https://github.com/apache/incubator-tvm/issues/4502) | |
| - add op crop_and_resize [#4417](https://github.com/apache/incubator-tvm/issues/4417) | |
| - Added pool autopadding and simplified converters. [#4672](https://github.com/apache/incubator-tvm/issues/4672) | |
| - Relay annotation and partitioning for external compilers [#4570](https://github.com/apache/incubator-tvm/issues/4570) | |
| - Fix meaning of conv2d_transpose output_padding parameter [#4318](https://github.com/apache/incubator-tvm/issues/4318) | |
| - use packed func macro for external codegen [#4710](https://github.com/apache/incubator-tvm/issues/4710) | |
| - fix _parse_param bug [#4711](https://github.com/apache/incubator-tvm/issues/4711) | |
| - Add constant input support for elemwise ops [#4666](https://github.com/apache/incubator-tvm/issues/4666) | |
| - Add parser support for squared difference [#4652](https://github.com/apache/incubator-tvm/issues/4652) | |
| - Add type check to dense [#4724](https://github.com/apache/incubator-tvm/issues/4724) | |
| - Invoke tvm::build from relay compile_engine and interpreter [#4723](https://github.com/apache/incubator-tvm/issues/4723) | |
| - Broadcast condition, x, and y for Where op [#4774](https://github.com/apache/incubator-tvm/issues/4774) | |
| - Add parser support for relational ops [#4695](https://github.com/apache/incubator-tvm/issues/4695) | |
| - Remove duplicated BindParamByName function in VM compiler [#4793](https://github.com/apache/incubator-tvm/issues/4793) | |
| - Use SimplifyInference for L2 Normalization. [#4795](https://github.com/apache/incubator-tvm/issues/4795) | |
| - Expose vm OptimizeModule to Python [#4800](https://github.com/apache/incubator-tvm/issues/4800) | |
| ### QNN | |
| - Making scale/zero_points as expr instead of attrs. [#4611](https://github.com/apache/incubator-tvm/issues/4611) | |
| - Channel wise quantization - Quantize & Requantize [#4629](https://github.com/apache/incubator-tvm/issues/4629) | |
| - Conv2D type checking for kernel per-channel scales. [#4732](https://github.com/apache/incubator-tvm/issues/4732) | |
| - Add missing nullptr check [#4773](https://github.com/apache/incubator-tvm/issues/4773) | |
| - Doc fix on convolution and dequantize [#4799](https://github.com/apache/incubator-tvm/issues/4799) | |
| - Conv2D with dilation support. [#4796](https://github.com/apache/incubator-tvm/issues/4796) | |
| ### VTA | |
| - Prevent Chisel VTA linter changing the Scala code [#4555](https://github.com/apache/incubator-tvm/issues/4555) | |
| - Update docker for TSIM based simulation [#4674](https://github.com/apache/incubator-tvm/issues/4674) | |
| - Update Jenkinsfile for VTA test with TSIM [#4734](https://github.com/apache/incubator-tvm/issues/4734) | |
| - Enable TSIM CI Testing [#4407](https://github.com/apache/incubator-tvm/issues/4407) | |
| - Fix an issue in updating uop_idx in the TensorGemm module [#4694](https://github.com/apache/incubator-tvm/issues/4694) | |
| - Support network which have no unique operator as start/stop name for graph pack. [#4703](https://github.com/apache/incubator-tvm/issues/4703) | |
| ### Runtime | |
| - make adt tag signed [#4605](https://github.com/apache/incubator-tvm/issues/4605) | |
| - Improve TVMBackendPackedCFunc to allow return val [#4637](https://github.com/apache/incubator-tvm/issues/4637) | |
| - EdgeTPU runtime for Coral Boards [#4698](https://github.com/apache/incubator-tvm/issues/4698) | |
| - Fix memory leak when using openMP [#4811](https://github.com/apache/incubator-tvm/issues/4811) | |
| ### TOPI | |
| - Allow empty tensor for reshape, tile and strided_slice [#4618](https://github.com/apache/incubator-tvm/issues/4618) | |
| - Fix meaning of conv2d_transpose output_padding parameter"; [#4708](https://github.com/apache/incubator-tvm/issues/4708) | |
| - Remove cpp upsampling and resize op [#4769](https://github.com/apache/incubator-tvm/issues/4769) | |
| - upsample operator 'NCHWinic' format support. [#4791](https://github.com/apache/incubator-tvm/issues/4791) | |
| ### CI | |
| - Pin python pillow to ";<7"; due to torchvision 1.2.0 dependency issue [#4632](https://github.com/apache/incubator-tvm/issues/4632) | |
| - Update image version tags in Dockerfile comments [#4631](https://github.com/apache/incubator-tvm/issues/4631) | |
| - better deletion script for pycache [#4635](https://github.com/apache/incubator-tvm/issues/4635) | |
| - Recover Windows Mac Build CI via Github Actions [#4662](https://github.com/apache/incubator-tvm/issues/4662) | |
| - Update deps for chisel [#4675](https://github.com/apache/incubator-tvm/issues/4675) | |
| - Bump to use the new cpu image [#4677](https://github.com/apache/incubator-tvm/issues/4677) | |
| ### Frontend | |
| - Add support for tf.Keras networks in Relay Keras frontend [#4630](https://github.com/apache/incubator-tvm/issues/4630) | |
| - Add conv3d [#4604](https://github.com/apache/incubator-tvm/issues/4604) | |
| - Fix incorrect calculations in tf SLICE [#4518](https://github.com/apache/incubator-tvm/issues/4518) | |
| ### Autotvm | |
| - Use VM compile to extract autotvm tasks [#4328](https://github.com/apache/incubator-tvm/issues/4328) | |
| - Download fallback schedule file if it does not exist [#4671](https://github.com/apache/incubator-tvm/issues/4671) | |
| - Ignore error when removing tmpdir [#4781](https://github.com/apache/incubator-tvm/issues/4781) | |
| - Fix a bug in generating the search space [#4779](https://github.com/apache/incubator-tvm/issues/4779) | |
| - Minor bug fixes in AutoTVM for QNN graphs [#4797](https://github.com/apache/incubator-tvm/issues/4797) | |
| ### Fixes | |
| - Make calibration faster and more memory usage friendly [#4589](https://github.com/apache/incubator-tvm/issues/4589) | |
| - Added declare of aluBits for TensorAlu [#4624](https://github.com/apache/incubator-tvm/issues/4624) | |
| - Improve comments [#4633](https://github.com/apache/incubator-tvm/issues/4633) | |
| - Get around limitation of g++-4.8 [#4626](https://github.com/apache/incubator-tvm/issues/4626) | |
| - Bugfix StmtMutator IfThenElse [#4609](https://github.com/apache/incubator-tvm/issues/4609) | |
| - Remove unecessary rdynamic [#4613](https://github.com/apache/incubator-tvm/issues/4613) | |
| - Resolve constexpr related link error in debug mode [#4641](https://github.com/apache/incubator-tvm/issues/4641) | |
| - Asymmetric padding [#4511](https://github.com/apache/incubator-tvm/issues/4511) | |
| - Reduce data size of asymmetric padding testcase [#4658](https://github.com/apache/incubator-tvm/issues/4658) | |
| - Fix Base64OutStream portability issue [#4668](https://github.com/apache/incubator-tvm/issues/4668) | |
| - fix topi.nn.global_pool layout="NHWC" [#4656](https://github.com/apache/incubator-tvm/issues/4656) | |
| - Also package core.rly [#4679](https://github.com/apache/incubator-tvm/issues/4679) | |
| - fskip of EliminateCommonSubexpr cannot always return false [#4620](https://github.com/apache/incubator-tvm/issues/4620) | |
| - Fix Python syntax error in start_rpc_server_to_tracker.py [#4682](https://github.com/apache/incubator-tvm/issues/4682) | |
| - os.path --> osp to match the import [#4681](https://github.com/apache/incubator-tvm/issues/4681) | |
| - GitHub actions/checkout@v1 --> v2 [#4680](https://github.com/apache/incubator-tvm/issues/4680) | |
| - Fix Python syntax error AGAIN in start_rpc_server_to_tracker.py [#4685](https://github.com/apache/incubator-tvm/issues/4685) | |
| - Use ==/!= to compare str, bytes, and int literals [#4686](https://github.com/apache/incubator-tvm/issues/4686) | |
| - Rename start_rpc_server_to_tracker.py to start_rpc_server_to_tracker.sh [#4689](https://github.com/apache/incubator-tvm/issues/4689) | |
| - Deploy Quantized Model on CUDA [#4667](https://github.com/apache/incubator-tvm/issues/4667) | |
| - Conv1D [#4639](https://github.com/apache/incubator-tvm/issues/4639) | |
| - 1D Pooling [#4663](https://github.com/apache/incubator-tvm/issues/4663) | |
| - GitHub Action lint Python code for syntax errors [#4688](https://github.com/apache/incubator-tvm/issues/4688) | |
| - Generate blob use LLVM directly [#4657](https://github.com/apache/incubator-tvm/issues/4657) | |
| - reduce input size to fix oom [#4653](https://github.com/apache/incubator-tvm/issues/4653) | |
| - Fix RemoveUnusedFunctions pass [#4700](https://github.com/apache/incubator-tvm/issues/4700) | |
| - link the math library by default [#4713](https://github.com/apache/incubator-tvm/issues/4713) | |
| - Update mainline version to 0.7.dev0 [#4720](https://github.com/apache/incubator-tvm/issues/4720) | |
| - add SizeVar representing non-neg valued variable in a tensor shape [#4684](https://github.com/apache/incubator-tvm/issues/4684) | |
| - Fix the compile problem of cpp_rpc [#4725](https://github.com/apache/incubator-tvm/issues/4725) | |
| - Bring Your Own Codegen Guide -- Part 1 [#4602](https://github.com/apache/incubator-tvm/issues/4602) | |
| - Convert Layout pass. [#4664](https://github.com/apache/incubator-tvm/issues/4664) | |
| - JSON upgrader to upgrade serialized json. [#4730](https://github.com/apache/incubator-tvm/issues/4730) | |
| - Fallback schedule for Int8 depthwise. [#4733](https://github.com/apache/incubator-tvm/issues/4733) | |
| - Fix dense x86 schedule [#4728](https://github.com/apache/incubator-tvm/issues/4728) | |
| - Fix demo dockerfile build failed [#4744](https://github.com/apache/incubator-tvm/issues/4744) | |
| - Expose relay BindParamsByName to Python [#4751](https://github.com/apache/incubator-tvm/issues/4751) | |
| - Improve CUDA vectorizer [#4736](https://github.com/apache/incubator-tvm/issues/4736) | |
| - Add .asf.yaml for github info [#4761](https://github.com/apache/incubator-tvm/issues/4761) | |
| - Bring Your Own Codegen Guide -- Part 2 [#4718](https://github.com/apache/incubator-tvm/issues/4718) | |
| - Fix padding in pooling op [#4738](https://github.com/apache/incubator-tvm/issues/4738) | |
| - Remove run_infer_type duplicates [#4766](https://github.com/apache/incubator-tvm/issues/4766) | |
| - pooling.cc improvements [#4767](https://github.com/apache/incubator-tvm/issues/4767) | |
| - Export builtin_fp16 on Windows [#4731](https://github.com/apache/incubator-tvm/issues/4731) | |
| - TVM_REGISTER_API -> TVM_REGISTER_GLOBAL [#4768](https://github.com/apache/incubator-tvm/issues/4768) | |
| - Fix Tensorflow conv3d pad bug, add non-cubic data and kernel tests [#4772](https://github.com/apache/incubator-tvm/issues/4772) | |
| - Bump prebuilt-image version in demo dockerfile [#4770](https://github.com/apache/incubator-tvm/issues/4770) | |
| - Update tune_simple_template.py [#4778](https://github.com/apache/incubator-tvm/issues/4778) | |
| - Explicitly link to cublasLt if it exists [#4776](https://github.com/apache/incubator-tvm/issues/4776) | |
| - Implement pass manager tracing API [#4782](https://github.com/apache/incubator-tvm/issues/4782) | |
| - Fix hasattr by extracting Python error type from Windows error message [#4780](https://github.com/apache/incubator-tvm/issues/4780) | |
| - Replace os.path.exists with try...except...else [#4784](https://github.com/apache/incubator-tvm/issues/4784) | |
| - Improve CUDA conv2d_transpose_nchw [#4762](https://github.com/apache/incubator-tvm/issues/4762) | |
| - Add CUDA conv2d for NHWC layout [#4737](https://github.com/apache/incubator-tvm/issues/4737) | |
| - Make sure to visit the arguments of inlined functions [#4783](https://github.com/apache/incubator-tvm/issues/4783) | |
| - Parse additional exception strings [#4785](https://github.com/apache/incubator-tvm/issues/4785) | |
| - conv3d_ndhwc schedule [#4775](https://github.com/apache/incubator-tvm/issues/4775) | |
| - fix #4670: add bias for fc layer [#4801](https://github.com/apache/incubator-tvm/issues/4801) | |
| - Change color channel from BGR to RGB for darknet preprocessing [#4794](https://github.com/apache/incubator-tvm/issues/4794) | |
| - Solve ARM BIG.LITTLE heterogeneous multicores [#4747](https://github.com/apache/incubator-tvm/issues/4747) | |
| - Create a StringImm reference type [#4806](https://github.com/apache/incubator-tvm/issues/4806) | |
| - Fix -Wextra [#4804](https://github.com/apache/incubator-tvm/issues/4804) | |
| - Fix vta tutorial [#4809](https://github.com/apache/incubator-tvm/issues/4809) | |
| ## Feb 2020 | |
| ### Fixes | |
| - conv3d_ndhwc schedule [#4775](https://github.com/apache/incubator-tvm/pull/4775) | |
| - fix #4670: add bias for fc layer [#4801](https://github.com/apache/incubator-tvm/pull/4801) | |
| - Change color channel from BGR to RGB for darknet preprocessing [#4794](https://github.com/apache/incubator-tvm/pull/4794) | |
| - Solve ARM BIG.LITTLE heterogeneous multicores [#4747](https://github.com/apache/incubator-tvm/pull/4747) | |
| - Create a StringImm reference type [#4806](https://github.com/apache/incubator-tvm/pull/4806) | |
| - Fix -Wextra [#4804](https://github.com/apache/incubator-tvm/pull/4804) | |
| - Minor bug fixes in AutoTVM for QNN graphs [#4797](https://github.com/apache/incubator-tvm/pull/4797) | |
| - Fixed subprocess creation under windows [#4820](https://github.com/apache/incubator-tvm/pull/4820) | |
| - It's gpu not cpu. [#4832](https://github.com/apache/incubator-tvm/pull/4832) | |
| - Improve tol to resolve flaky case [#4836](https://github.com/apache/incubator-tvm/pull/4836) | |
| - comaniac -> reviewer [#4841](https://github.com/apache/incubator-tvm/pull/4841) | |
| - Fixed process termination routine in windows [#4844](https://github.com/apache/incubator-tvm/pull/4844) | |
| - test_cuddn flaky [#4846](https://github.com/apache/incubator-tvm/pull/4846) | |
| - Mxnet parser for Qnn dialect [#4714](https://github.com/apache/incubator-tvm/pull/4714) | |
| - Enhance cc.cross_compiler [#4817](https://github.com/apache/incubator-tvm/pull/4817) | |
| - Fixed crash caused by reversing bitwise operations [#4852](https://github.com/apache/incubator-tvm/pull/4852) | |
| - reverse some changes made for intel_graphics/conv2d.py in PR #4849 [#4853](https://github.com/apache/incubator-tvm/pull/4853) | |
| - const auto p -> const auto& p [#4861](https://github.com/apache/incubator-tvm/pull/4861) | |
| - Fix onnx import bugs [#4750](https://github.com/apache/incubator-tvm/pull/4750) | |
| - Explicit llvm::StringRef to std::string conversion [#4859](https://github.com/apache/incubator-tvm/pull/4859) | |
| - Optimize x86 conv3d_ndhwc using data packing approach. [#4866](https://github.com/apache/incubator-tvm/pull/4866) | |
| - Update the runtime PackedFunc for module [#4871](https://github.com/apache/incubator-tvm/pull/4871) | |
| - improve antlr import error message [#4888](https://github.com/apache/incubator-tvm/pull/4888) | |
| - Update faq.md [#4893](https://github.com/apache/incubator-tvm/pull/4893) | |
| - Fix alpha_equal bug for attribute check [#4897](https://github.com/apache/incubator-tvm/pull/4897) | |
| - Fix issues in cuda codegen [#4876](https://github.com/apache/incubator-tvm/pull/4876) | |
| - Support range in index based tuners [#4870](https://github.com/apache/incubator-tvm/pull/4870) | |
| - Fast exponent [#4790](https://github.com/apache/incubator-tvm/pull/4790) | |
| - Fixed: Bitwise ops on floats causing wrong code generation and crashes. [#4892](https://github.com/apache/incubator-tvm/pull/4892) | |
| - Fix tvm.target.generic_func runtime detection [#4910](https://github.com/apache/incubator-tvm/pull/4910) | |
| - Support cuda tensorcore subbyte int data type in auto tensorcore [#4546](https://github.com/apache/incubator-tvm/pull/4546) | |
| - @anijain2305 -> Committer [#4921](https://github.com/apache/incubator-tvm/pull/4921) | |
| - topi/tests/python/test_topi_sort.py::test_argsort [#4891](https://github.com/apache/incubator-tvm/pull/4891) | |
| - Use opencv reisze method for preprocessing of image in darknet [#4883](https://github.com/apache/incubator-tvm/pull/4883) | |
| - Fix build breaks with StringRef changes [#4923](https://github.com/apache/incubator-tvm/pull/4923) | |
| - remove unnecessary spliting in the cached chunk [#4935](https://github.com/apache/incubator-tvm/pull/4935) | |
| - Fixing an Infinite Loop case in UnmatchedChecker. [#4881](https://github.com/apache/incubator-tvm/pull/4881) | |
| - Tensor Expression Debug Display (TEDD) [#4651](https://github.com/apache/incubator-tvm/pull/4651) | |
| - Add a tutorial for PyTorch [#4936](https://github.com/apache/incubator-tvm/pull/4936) | |
| - Remove SGX toolchain installation from CI Dockerfile [#4948](https://github.com/apache/incubator-tvm/pull/4948) | |
| - YoloV3 Support [#4887](https://github.com/apache/incubator-tvm/pull/4887) | |
| - Fix tedd tutorial after strategy change [#4947](https://github.com/apache/incubator-tvm/pull/4947) | |
| - Bump up dev version [#4941](https://github.com/apache/incubator-tvm/pull/4941) | |
| - Allow customize MKLDNN library location [#4814](https://github.com/apache/incubator-tvm/pull/4814) | |
| - Move Ops in relay.op.contrib.* [#4942](https://github.com/apache/incubator-tvm/pull/4942) | |
| - Added CopyFromBytes and CopyToBytes convenience methods to NDArray. Fixed typos. [#4970](https://github.com/apache/incubator-tvm/pull/4970) | |
| - Fix gcn tutorial failure [#4994](https://github.com/apache/incubator-tvm/pull/4994) | |
| - @huajsj -> Reviewer [#4993](https://github.com/apache/incubator-tvm/pull/4993) | |
| - Fix stride default value None in torch.nn.functional.avg_pool [#4984](https://github.com/apache/incubator-tvm/pull/4984) | |
| - fix ROCm strategy for winograd conv selection [#5001](https://github.com/apache/incubator-tvm/pull/5001) | |
| - Fix get_valid_count flaky test for cuda [#4901](https://github.com/apache/incubator-tvm/pull/4901) | |
| - @optima2005 -> reviewer [#5004](https://github.com/apache/incubator-tvm/pull/5004) | |
| - Change Scala Linter scalafmt => scalastyle [#4998](https://github.com/apache/incubator-tvm/pull/4998) | |
| - kill from tvm import te [#5007](https://github.com/apache/incubator-tvm/pull/5007) | |
| - Docs and Readme updated as per new namespace change [#4989](https://github.com/apache/incubator-tvm/pull/4989) | |
| - Chisel fixes and de10nano support [#4986](https://github.com/apache/incubator-tvm/pull/4986) | |
| - Conditions updated to cover better user scenarios [#4951](https://github.com/apache/incubator-tvm/pull/4951) | |
| - Tighten split's extent [#4931](https://github.com/apache/incubator-tvm/pull/4931) | |
| - Fix gpu not found when running TVM docker [#4975](https://github.com/apache/incubator-tvm/pull/4975) | |
| - typo [#5008](https://github.com/apache/incubator-tvm/pull/5008) | |
| - Fixes for pylint==2.4.4 [#4849](https://github.com/apache/incubator-tvm/pull/4849) | |
| - Add docs for tvm.ir [#4869](https://github.com/apache/incubator-tvm/pull/4869) | |
| ### Relay | |
| - Expose vm OptimizeModule to Python [#4800](https://github.com/apache/incubator-tvm/pull/4800) | |
| - Add parser support for logical operators [#4642](https://github.com/apache/incubator-tvm/pull/4642) | |
| - Conv2D padding representation [#4787](https://github.com/apache/incubator-tvm/pull/4787) | |
| - Add bitwise ops [#4815](https://github.com/apache/incubator-tvm/pull/4815) | |
| - Add support for quantized LOGISTIC [#4696](https://github.com/apache/incubator-tvm/pull/4696) | |
| - Added Merge Composite pass [#4771](https://github.com/apache/incubator-tvm/pull/4771) | |
| - Fix VM compiler for while loop with free vars [#4889](https://github.com/apache/incubator-tvm/pull/4889) | |
| - Fix bug in re-processing call node in MergeComposite pass [#4879](https://github.com/apache/incubator-tvm/pull/4879) | |
| - NHWC import support. [#4899](https://github.com/apache/incubator-tvm/pull/4899) | |
| - Expose FunctionGetAttr to Python [#4905](https://github.com/apache/incubator-tvm/pull/4905) | |
| - Fix FuseBatchNorm output cast error if need_cast is True [#4894](https://github.com/apache/incubator-tvm/pull/4894) | |
| - Fix an assertion exposed by loop vectorizer [#4916](https://github.com/apache/incubator-tvm/pull/4916) | |
| - Relay op strategy [#4644](https://github.com/apache/incubator-tvm/pull/4644) | |
| - Add a PyTorch to Relay Parser [#4497](https://github.com/apache/incubator-tvm/pull/4497) | |
| - Support data types for CSourceModuleCodegen args and output [#4934](https://github.com/apache/incubator-tvm/pull/4934) | |
| - fix error message [#4945](https://github.com/apache/incubator-tvm/pull/4945) | |
| - Clean up and refactor PyTorch frontend [#4944](https://github.com/apache/incubator-tvm/pull/4944) | |
| - Relay pass to use fast exp/tanh [#4873](https://github.com/apache/incubator-tvm/pull/4873) | |
| - call graph for relay [#4922](https://github.com/apache/incubator-tvm/pull/4922) | |
| - Add inline pass [#4927](https://github.com/apache/incubator-tvm/pull/4927) | |
| - Target annotation for external codegen [#4933](https://github.com/apache/incubator-tvm/pull/4933) | |
| - BatchNorm support with run-time mean and variance calculation [#4990](https://github.com/apache/incubator-tvm/pull/4990) | |
| - Reduce plevel of conv2d winograd implementation on cuda [#4987](https://github.com/apache/incubator-tvm/pull/4987) | |
| - add operation tan to TVM [#4938](https://github.com/apache/incubator-tvm/pull/4938) | |
| - outline and inline lifted functions for external codegen [#4996](https://github.com/apache/incubator-tvm/pull/4996) | |
| ### Qnn | |
| - Doc fix on convolution and dequantize [#4799](https://github.com/apache/incubator-tvm/pull/4799) | |
| - Conv2D with dilation support. [#4796](https://github.com/apache/incubator-tvm/pull/4796) | |
| - Optimize lowering for requantize and FixedPointMultiply. [#4798](https://github.com/apache/incubator-tvm/pull/4798) | |
| - More doc fix on quantize and convolution [#4874](https://github.com/apache/incubator-tvm/pull/4874) | |
| - Add support for per channel weight scale in dense op [#4880](https://github.com/apache/incubator-tvm/pull/4880) | |
| - Add support for quantized models via QNN [#4977](https://github.com/apache/incubator-tvm/pull/4977) | |
| - Add support for quantized models via QNN (#4977)" [#5013](https://github.com/apache/incubator-tvm/pull/5013) | |
| ### Topi | |
| - upsample operator 'NCHWinic' format support. [#4791](https://github.com/apache/incubator-tvm/pull/4791) | |
| - Injective schedule improvement [#4786](https://github.com/apache/incubator-tvm/pull/4786) | |
| - Missing header [#4865](https://github.com/apache/incubator-tvm/pull/4865) | |
| - Enable vectorization on fp16 type [#4867](https://github.com/apache/incubator-tvm/pull/4867) | |
| - fix docs errors [#4973](https://github.com/apache/incubator-tvm/pull/4973) | |
| ### Refactor | |
| - tvm._ffi [#4813](https://github.com/apache/incubator-tvm/pull/4813) | |
| - Establish tvm.runtime [#4818](https://github.com/apache/incubator-tvm/pull/4818) | |
| - move vm.py under runtime and adt to runtime.container.py [#4855](https://github.com/apache/incubator-tvm/pull/4855) | |
| - establish tvm.ir, migrate corresponding files [#4862](https://github.com/apache/incubator-tvm/pull/4862) | |
| - Polish tvm.runtime, tvm.runtime.module API update [#4837](https://github.com/apache/incubator-tvm/pull/4837) | |
| - Establish tvm.target [#4872](https://github.com/apache/incubator-tvm/pull/4872) | |
| - Establish tvm.tir [#4877](https://github.com/apache/incubator-tvm/pull/4877) | |
| - Establish tvm.te and tvm.driver [#4900](https://github.com/apache/incubator-tvm/pull/4900) | |
| - Establish tvm.arith [#4904](https://github.com/apache/incubator-tvm/pull/4904) | |
| - Polish ffi convention. [#4912](https://github.com/apache/incubator-tvm/pull/4912) | |
| - Remove legacy python files. [#4943](https://github.com/apache/incubator-tvm/pull/4943) | |
| - refactor build module to take IRModule [#4988](https://github.com/apache/incubator-tvm/pull/4988) | |
| ### Frontend | |
| - Dynamically calculate input_stats of any fake_quant range [#4789](https://github.com/apache/incubator-tvm/pull/4789) | |
| - LSTM Support [#4825](https://github.com/apache/incubator-tvm/pull/4825) | |
| - Add MIRROR_PAD operator [#4822](https://github.com/apache/incubator-tvm/pull/4822) | |
| - use qnn helper function in softmax [#4840](https://github.com/apache/incubator-tvm/pull/4840) | |
| - Add Resize op converter [#4838](https://github.com/apache/incubator-tvm/pull/4838) | |
| - Add support for TFLite_Detection_PostProcess [#4543](https://github.com/apache/incubator-tvm/pull/4543) | |
| - Fix tests for tflite unary elemwise operations [#4913](https://github.com/apache/incubator-tvm/pull/4913) | |
| - GaussianDropout/Noise parsing support [#4928](https://github.com/apache/incubator-tvm/pull/4928) | |
| - Add parser support for 'square' operator [#4915](https://github.com/apache/incubator-tvm/pull/4915) | |
| - make_loss operator support [#4930](https://github.com/apache/incubator-tvm/pull/4930) | |
| - Add parser support for l2_normalization [#4966](https://github.com/apache/incubator-tvm/pull/4966) | |
| - ReadVariableOp operator support [#4952](https://github.com/apache/incubator-tvm/pull/4952) | |
| - Check graph inputs match expected [#4992](https://github.com/apache/incubator-tvm/pull/4992) | |
| - support multiply outputs [#4980](https://github.com/apache/incubator-tvm/pull/4980) | |
| ### Ci | |
| - Update ci-gpu to v0.60 [#4827](https://github.com/apache/incubator-tvm/pull/4827) | |
| - Update ci-gpu torch1.4 and onnx1.6 [#4826](https://github.com/apache/incubator-tvm/pull/4826) | |
| - Update ci-lint to v0.60 [#4850](https://github.com/apache/incubator-tvm/pull/4850) | |
| - Update ci-lint to pylint2.4.4 [#4851](https://github.com/apache/incubator-tvm/pull/4851) | |
| - Cleanup logfile before tutorial runs [#4896](https://github.com/apache/incubator-tvm/pull/4896) | |
| - Update ci docker to add autodocsumm [#4903](https://github.com/apache/incubator-tvm/pull/4903) | |
| - Add autodocsum as dep [#4902](https://github.com/apache/incubator-tvm/pull/4902) | |
| - Add pre-check script to check sphinx doc build. [#4956](https://github.com/apache/incubator-tvm/pull/4956) | |
| - pin xgboost dependency version to 0.90 [#4965](https://github.com/apache/incubator-tvm/pull/4965) | |
| ### Doc | |
| - Fix doc after moving to unified IR [#4835](https://github.com/apache/incubator-tvm/pull/4835) | |
| - Introduction to module serialization [#4564](https://github.com/apache/incubator-tvm/pull/4564) | |
| - ConvertLayout - Call RemoveUnunsedFunctions. [#4834](https://github.com/apache/incubator-tvm/pull/4834) | |
| - Fix bugs that override n_trials [#4842](https://github.com/apache/incubator-tvm/pull/4842) | |
| - Update the vm doc [#4868](https://github.com/apache/incubator-tvm/pull/4868) | |
| - Refine the example description of max/min/sum/tag_scope [#4974](https://github.com/apache/incubator-tvm/pull/4974) | |
| ### Docs | |
| - Fix vta tutorial [#4809](https://github.com/apache/incubator-tvm/pull/4809) | |
| - Introduce how to add hardware backend to FAQ [#4898](https://github.com/apache/incubator-tvm/pull/4898) | |
| - Update API docs to reflect the status after the refactor. [#4907](https://github.com/apache/incubator-tvm/pull/4907) | |
| - Fix sphinx warnings [#4917](https://github.com/apache/incubator-tvm/pull/4917) | |
| - Fix Sphinx Warnings (RST indent, cross-ref, and image scale) [#4920](https://github.com/apache/incubator-tvm/pull/4920) | |
| - Fix Sphinx Warning: the target found for cross-reference [#4925](https://github.com/apache/incubator-tvm/pull/4925) | |
| - Sphinx -- Introduce alias detection. [#4954](https://github.com/apache/incubator-tvm/pull/4954) | |
| - Fix Warnings from #4942 [#4959](https://github.com/apache/incubator-tvm/pull/4959) | |
| - Fix sphinx precheck [#4967](https://github.com/apache/incubator-tvm/pull/4967) | |
| ### Runtime | |
| - Fix memory leakage of TVMByteArray [#4856](https://github.com/apache/incubator-tvm/pull/4856) | |
| - Fix TVM_DLL_EXPORT_TYPED_FUNC to work on Windows [#4955](https://github.com/apache/incubator-tvm/pull/4955) | |
| - Fix memory leak when using openMP [#4811](https://github.com/apache/incubator-tvm/pull/4811) | |
| - Export GraphRuntime in tvm_runtime.dll [#5002](https://github.com/apache/incubator-tvm/pull/5002) | |
| - MISRA-C compliant TVM runtime [#3934](https://github.com/apache/incubator-tvm/pull/3934) | |
| ### Tflite | |
| - Using real image for QNN testing. [#4816](https://github.com/apache/incubator-tvm/pull/4816) | |
| - FLOOR_MOD & FLOOR_DIV support [#4971](https://github.com/apache/incubator-tvm/pull/4971) | |
| ### Torch | |
| - Upsampling op support and enable registering a user defined op conversion map [#4961](https://github.com/apache/incubator-tvm/pull/4961) | |
| - fix unordered dictionary problem for python version under 3.6 [#4982](https://github.com/apache/incubator-tvm/pull/4982) | |
| ## Mar 2020 | |
| ### IR | |
| - Introduce tir::PrimFunc [#5070](https://github.com/apache/incubator-tvm/pull/5070) | |
| - Update the type_keys to reflect the code-org [#5074](https://github.com/apache/incubator-tvm/pull/5074) | |
| - Fix AttrEqual for Array and StrMap, double [#5054](https://github.com/apache/incubator-tvm/pull/5054) | |
| - Introduce StructuralHash for the Unified IR. [#5160](https://github.com/apache/incubator-tvm/pull/5160) | |
| - Introduce StructuralEqual Infra for the unified IR. [#5154](https://github.com/apache/incubator-tvm/pull/5154) | |
| - Add dump and print for debugging (NFC) [#5207](https://github.com/apache/incubator-tvm/pull/5207) | |
| ### Relay | |
| - Relay pass to use fast exp/tanh [#4873](https://github.com/apache/incubator-tvm/pull/4873) | |
| - Add inline pass [#4927](https://github.com/apache/incubator-tvm/pull/4927) | |
| - Target annotation for external codegen [#4933](https://github.com/apache/incubator-tvm/pull/4933) | |
| - BatchNorm support with run-time mean and variance calculation [#4990](https://github.com/apache/incubator-tvm/pull/4990) | |
| - add operation tan to TVM [#4938](https://github.com/apache/incubator-tvm/pull/4938) | |
| - outline and inline lifted functions for external codegen [#4996](https://github.com/apache/incubator-tvm/pull/4996) | |
| - Reduce plevel of conv2d winograd implementation on cuda [#4987](https://github.com/apache/incubator-tvm/pull/4987) | |
| - Remove primitive attribute from composite function [#5014](https://github.com/apache/incubator-tvm/pull/5014) | |
| - Add submodule extraction pass [#4960](https://github.com/apache/incubator-tvm/pull/4960) | |
| - Refactor Adaptive pool and add 3d support [#5049](https://github.com/apache/incubator-tvm/pull/5049) | |
| - Support for Int8 schedules - CUDA/x86 [#5031](https://github.com/apache/incubator-tvm/pull/5031) | |
| - Dilation2D operator support [#5033](https://github.com/apache/incubator-tvm/pull/5033) | |
| - Refactor Relay Python to use new FFI [#5077](https://github.com/apache/incubator-tvm/pull/5077) | |
| - Fix relay node registration after refactor [#5083](https://github.com/apache/incubator-tvm/pull/5083) | |
| - Codegen_c.h should include relay.function [#5093](https://github.com/apache/incubator-tvm/pull/5093) | |
| - Move expr.Function to function.py [#5087](https://github.com/apache/incubator-tvm/pull/5087) | |
| - operator support NonZero [#5073](https://github.com/apache/incubator-tvm/pull/5073) | |
| - new PR to re-add tan to TVM [#5025](https://github.com/apache/incubator-tvm/pull/5025) | |
| - Propagate constant to subgraphs [#5094](https://github.com/apache/incubator-tvm/pull/5094) | |
| - Adjust strategy plevel to achieve expected performance by default [#5118](https://github.com/apache/incubator-tvm/pull/5118) | |
| - Isfinite operator [#4981](https://github.com/apache/incubator-tvm/pull/4981) | |
| - Unravel Index operator [#5082](https://github.com/apache/incubator-tvm/pull/5082) | |
| - add operation tan to TVM (#4938)" [#5017](https://github.com/apache/incubator-tvm/pull/5017) | |
| - GradientCell Relay Pass [#5039](https://github.com/apache/incubator-tvm/pull/5039) | |
| - Added a AnnotatedRegion utility class [#5030](https://github.com/apache/incubator-tvm/pull/5030) | |
| - Register topi schedule for Relay fast_exp and fast_tanh [#5131](https://github.com/apache/incubator-tvm/pull/5131) | |
| - Support TupleGetItem in body of pattern [#5106](https://github.com/apache/incubator-tvm/pull/5106) | |
| - Add MergeCompilerRegions pass [#5134](https://github.com/apache/incubator-tvm/pull/5134) | |
| - Partition graph codestyle fixes [#5202](https://github.com/apache/incubator-tvm/pull/5202) | |
| - Re-wrote the Graph Partitioner to support multiple outputs [#5143](https://github.com/apache/incubator-tvm/pull/5143) | |
| - Fixes to MergeCompilerRegions [#5195](https://github.com/apache/incubator-tvm/pull/5195) | |
| ### Torch | |
| - Upsampling op support and enable registering a user defined op conversion map [#4961](https://github.com/apache/incubator-tvm/pull/4961) | |
| - fix unordered dictionary problem for python version under 3.6 [#4982](https://github.com/apache/incubator-tvm/pull/4982) | |
| - Check graph inputs match expected [#4992](https://github.com/apache/incubator-tvm/pull/4992) | |
| - Add support for quantized models via QNN [#4977](https://github.com/apache/incubator-tvm/pull/4977) | |
| - Add initial control flow support [#4964](https://github.com/apache/incubator-tvm/pull/4964) | |
| - Remove FP32 piggy back and use QNN add/mul/concatenate [#5061](https://github.com/apache/incubator-tvm/pull/5061) | |
| - Add missing upcast to uint8 avg_pool conversion [#5089](https://github.com/apache/incubator-tvm/pull/5089) | |
| - Add initial 3D op support and test on Resnet 3D [#5075](https://github.com/apache/incubator-tvm/pull/5075) | |
| - Fix conv2d conversion for group conv (group > 1 but != in channels) [#5132](https://github.com/apache/incubator-tvm/pull/5132) | |
| - Add support for max_pool1d [#5142](https://github.com/apache/incubator-tvm/pull/5142) | |
| - Add support for split [#5174](https://github.com/apache/incubator-tvm/pull/5174) | |
| ### Tflite | |
| - FLOOR_MOD & FLOOR_DIV support [#4971](https://github.com/apache/incubator-tvm/pull/4971) | |
| - Activation functions support [#4978](https://github.com/apache/incubator-tvm/pull/4978) | |
| - Round op parsing support added [#5022](https://github.com/apache/incubator-tvm/pull/5022) | |
| - DepthToSpace and SpaceToDepth support [#5041](https://github.com/apache/incubator-tvm/pull/5041) | |
| - TOP_K op parser support [#5051](https://github.com/apache/incubator-tvm/pull/5051) | |
| ### Frontend | |
| - ReadVariableOp operator support [#4952](https://github.com/apache/incubator-tvm/pull/4952) | |
| - support multiply outputs [#4980](https://github.com/apache/incubator-tvm/pull/4980) | |
| - Reduce_any op parsing support [#4926](https://github.com/apache/incubator-tvm/pull/4926) | |
| - TensorFlow Parser Control Flow Enhancement [#5020](https://github.com/apache/incubator-tvm/pull/5020) | |
| - TensorFlow Frontend support with shared params [#5042](https://github.com/apache/incubator-tvm/pull/5042) | |
| - Support for AddV2 in Relay Tensorflow frontend converter. [#5046](https://github.com/apache/incubator-tvm/pull/5046) | |
| - conv3d frontend operator support [#5080](https://github.com/apache/incubator-tvm/pull/5080) | |
| - Max_pool3d and Averagepool3d operator support [#5085](https://github.com/apache/incubator-tvm/pull/5085) | |
| - Support for Atan/Atan2 in Relay Tensorflow frontend converter. [#5104](https://github.com/apache/incubator-tvm/pull/5104) | |
| - Use leaky by default for LeakyReLU [#5192](https://github.com/apache/incubator-tvm/pull/5192) | |
| ### Topi | |
| - Move Dilation2d from nn to image namespace [#5110](https://github.com/apache/incubator-tvm/pull/5110) | |
| - Use Thrust sort for argsort and topk [#5097](https://github.com/apache/incubator-tvm/pull/5097) | |
| - Conv2d and Dense ops support on Tensor Core [#5099](https://github.com/apache/incubator-tvm/pull/5099) | |
| - Setting workload correctly for Depthwise Spatial conv ARM. [#5182](https://github.com/apache/incubator-tvm/pull/5182) | |
| ### CI | |
| - pin xgboost dependency version to 0.90 [#4965](https://github.com/apache/incubator-tvm/pull/4965) | |
| - Temporary disable rust test [#5029](https://github.com/apache/incubator-tvm/pull/5029) | |
| - Change Rust version to stable in Docker [#5138](https://github.com/apache/incubator-tvm/pull/5138) | |
| - Update rust docker [#5141](https://github.com/apache/incubator-tvm/pull/5141) | |
| - Move build configuration to shell scripts [#5164](https://github.com/apache/incubator-tvm/pull/5164) | |
| - Improve VTA build message and scripts. [#5170](https://github.com/apache/incubator-tvm/pull/5170) | |
| ### Refactor | |
| - refactor build module to take IRModule [#4988](https://github.com/apache/incubator-tvm/pull/4988) | |
| - Streamline Function Attr interface. [#5045](https://github.com/apache/incubator-tvm/pull/5045) | |
| - Separate analysis and transform passes [#5035](https://github.com/apache/incubator-tvm/pull/5035) | |
| - Make target codegen take IRModule and PrimFunc. [#5107](https://github.com/apache/incubator-tvm/pull/5107) | |
| - Relay Node::make to constructor [#5128](https://github.com/apache/incubator-tvm/pull/5128) | |
| - Introduce PrimFuncPass. [#5139](https://github.com/apache/incubator-tvm/pull/5139) | |
| - Introducing VTA_HW_PATH for easier migration [#5163](https://github.com/apache/incubator-tvm/pull/5163) | |
| - alpha_equal to structural_equal [#5161](https://github.com/apache/incubator-tvm/pull/5161) | |
| - relay::StructuralHash to tvm::StructuralHash [#5166](https://github.com/apache/incubator-tvm/pull/5166) | |
| - Remove AttrsEqual and AttrsHash related code [#5169](https://github.com/apache/incubator-tvm/pull/5169) | |
| - Migrate Low-level Passes to Pass Manager [#5198](https://github.com/apache/incubator-tvm/pull/5198) | |
| - Introduce ExprDeepEqual, Remove IRDeepCompare [#5206](https://github.com/apache/incubator-tvm/pull/5206) | |
| ### Runtime | |
| - Export GraphRuntime in tvm_runtime.dll [#5002](https://github.com/apache/incubator-tvm/pull/5002) | |
| - fix unused-value warning [#5140](https://github.com/apache/incubator-tvm/pull/5140) | |
| - crt error handling [#5147](https://github.com/apache/incubator-tvm/pull/5147) | |
| - MISRA-C compliant TVM runtime [#3934](https://github.com/apache/incubator-tvm/pull/3934) | |
| - Bundle deployment with static linking [#5158](https://github.com/apache/incubator-tvm/pull/5158) | |
| ### Vta | |
| - Change Scala Linter scalafmt => scalastyle [#4998](https://github.com/apache/incubator-tvm/pull/4998) | |
| - Chisel fixes and de10nano support [#4986](https://github.com/apache/incubator-tvm/pull/4986) | |
| - VTA hardware/software codebase re-org [#5037](https://github.com/apache/incubator-tvm/pull/5037) | |
| - HW sources refactor [#5188](https://github.com/apache/incubator-tvm/pull/5188) | |
| ### Qnn | |
| - Add support for quantized models via QNN (#4977)" [#5013](https://github.com/apache/incubator-tvm/pull/5013) | |
| - Support 4D padding. [#5036](https://github.com/apache/incubator-tvm/pull/5036) | |
| ### Docs | |
| - Move git_howto to rst, add Stage documents to te [#5055](https://github.com/apache/incubator-tvm/pull/5055) | |
| - Add doc for Relay op strategy [#5078](https://github.com/apache/incubator-tvm/pull/5078) | |
| - Update relay docs [#5112](https://github.com/apache/incubator-tvm/pull/5112) | |
| - include a tarball of docs, add a security faq [#5119](https://github.com/apache/incubator-tvm/pull/5119) | |
| - Cleanup docs before rebuild [#5127](https://github.com/apache/incubator-tvm/pull/5127) | |
| - Minimize necessary doc change [#5129](https://github.com/apache/incubator-tvm/pull/5129) | |
| - Various sphinx related fix. [#5168](https://github.com/apache/incubator-tvm/pull/5168) | |
| - Point docs to the ASF site. [#5178](https://github.com/apache/incubator-tvm/pull/5178) | |
| - Use https link [#5183](https://github.com/apache/incubator-tvm/pull/5183) | |
| - Reduce artifcats generated by sphinx gallery [#5208](https://github.com/apache/incubator-tvm/pull/5208) | |
| ### Fixes | |
| - fix docs errors [#4973](https://github.com/apache/incubator-tvm/pull/4973) | |
| - Refine the example description of max/min/sum/tag_scope [#4974](https://github.com/apache/incubator-tvm/pull/4974) | |
| - Fix gcn tutorial failure [#4994](https://github.com/apache/incubator-tvm/pull/4994) | |
| - @huajsj -> Reviewer [#4993](https://github.com/apache/incubator-tvm/pull/4993) | |
| - Fix stride default value None in torch.nn.functional.avg_pool [#4984](https://github.com/apache/incubator-tvm/pull/4984) | |
| - fix ROCm strategy for winograd conv selection [#5001](https://github.com/apache/incubator-tvm/pull/5001) | |
| - @optima2005 -> reviewer [#5004](https://github.com/apache/incubator-tvm/pull/5004) | |
| - kill from tvm import te [#5007](https://github.com/apache/incubator-tvm/pull/5007) | |
| - Docs and Readme updated as per new namespace change [#4989](https://github.com/apache/incubator-tvm/pull/4989) | |
| - Fix gpu not found when running TVM docker [#4975](https://github.com/apache/incubator-tvm/pull/4975) | |
| - typo [#5008](https://github.com/apache/incubator-tvm/pull/5008) | |
| - Add support for quantized models via QNN [#5016](https://github.com/apache/incubator-tvm/pull/5016) | |
| - Revive the Rust + SGX refactor [#4976](https://github.com/apache/incubator-tvm/pull/4976) | |
| - Implemented kDLCPUPinned (cudaMallocHost) [#4985](https://github.com/apache/incubator-tvm/pull/4985) | |
| - Early checking added and new test cases added for schedule fuse [#5010](https://github.com/apache/incubator-tvm/pull/5010) | |
| - Tighten split's extent [#4931](https://github.com/apache/incubator-tvm/pull/4931) | |
| - Adding a few missing math intrin [#5011](https://github.com/apache/incubator-tvm/pull/5011) | |
| - Revert "Conditions updated to cover better user scenarios" [#5032](https://github.com/apache/incubator-tvm/pull/5032) | |
| - Revert "Tighten split's extent" [#5027](https://github.com/apache/incubator-tvm/pull/5027) | |
| - Conditions updated to cover better user scenarios [#4951](https://github.com/apache/incubator-tvm/pull/4951) | |
| - Conv3D ONNX support and conv3D_ncdhw x86 schedules [#4949](https://github.com/apache/incubator-tvm/pull/4949) | |
| - Conditions updated to cover better user scenarios [#5043](https://github.com/apache/incubator-tvm/pull/5043) | |
| - Add String container [#4628](https://github.com/apache/incubator-tvm/pull/4628) | |
| - Set split node's range to minimum of ext and split factor or split np… [#5044](https://github.com/apache/incubator-tvm/pull/5044) | |
| - Fixed div by zero core dump. Fixed rounding intrinsics on int crash [#5026](https://github.com/apache/incubator-tvm/pull/5026) | |
| - Test case modified for int type [#5012](https://github.com/apache/incubator-tvm/pull/5012) | |
| - CI: Install apt-transport-https [#5053](https://github.com/apache/incubator-tvm/pull/5053) | |
| - Fix autotvm customized template [#5034](https://github.com/apache/incubator-tvm/pull/5034) | |
| - fix typo [#5058](https://github.com/apache/incubator-tvm/pull/5058) | |
| - Add opt out operator for has_multiple_inputs for graph tuner [#5000](https://github.com/apache/incubator-tvm/pull/5000) | |
| - Bug Fix for ARM CPUs. Lower strict assumption. [#5063](https://github.com/apache/incubator-tvm/pull/5063) | |
| - Add support for FusedBatchNormV3 [#5065](https://github.com/apache/incubator-tvm/pull/5065) | |
| - Triage the testcases to fit the the new namespaces [#5071](https://github.com/apache/incubator-tvm/pull/5071) | |
| - Require python3.6 [#5057](https://github.com/apache/incubator-tvm/pull/5057) | |
| - Require c++14 by default [#5056](https://github.com/apache/incubator-tvm/pull/5056) | |
| - Tensorflow/lite scripts upgrade from 1.13.1 to 2.0.0 [#4963](https://github.com/apache/incubator-tvm/pull/4963) | |
| - Change Azure pipeline badge to GH actions [#5081](https://github.com/apache/incubator-tvm/pull/5081) | |
| - Explicitly cast min/max operands [#5090](https://github.com/apache/incubator-tvm/pull/5090) | |
| - Description updated for pooling attributes [#5091](https://github.com/apache/incubator-tvm/pull/5091) | |
| - Fix vta tutorial after relay function refactor [#5095](https://github.com/apache/incubator-tvm/pull/5095) | |
| - Support QNN ops. [#5066](https://github.com/apache/incubator-tvm/pull/5066) | |
| - Enhance CUDA codegen for SelectNode [#4983](https://github.com/apache/incubator-tvm/pull/4983) | |
| - Replace UseDefaultCompiler with GetAttr [#5088](https://github.com/apache/incubator-tvm/pull/5088) | |
| - Add colors to compute_at edges and thread/block indices. [#5111](https://github.com/apache/incubator-tvm/pull/5111) | |
| - Temporary fix to the stack overflow issue in autotvm task extraction [#5019](https://github.com/apache/incubator-tvm/pull/5019) | |
| - Fix compilation of If-Elses [#5040](https://github.com/apache/incubator-tvm/pull/5040) | |
| - Fix CompilerAttrs [#5109](https://github.com/apache/incubator-tvm/pull/5109) | |
| - TVM android camera demo [#5005](https://github.com/apache/incubator-tvm/pull/5005) | |
| - Update the tarball deployment. [#5120](https://github.com/apache/incubator-tvm/pull/5120) | |
| - Fix the existing test cases before refactoring. [#5122](https://github.com/apache/incubator-tvm/pull/5122) | |
| - Fix typo for deploy_model_on_android.py [#5123](https://github.com/apache/incubator-tvm/pull/5123) | |
| - Vectorization for intrinsics [#5101](https://github.com/apache/incubator-tvm/pull/5101) | |
| - Fixed bug where shifting by out-of-bounds value results in no compute code being emitted. [#5115](https://github.com/apache/incubator-tvm/pull/5115) | |
| - Add thrust support for nms [#5116](https://github.com/apache/incubator-tvm/pull/5116) | |
| - Fix for issue #4831. The data_min_idx and data_max_idx were flipped. [#5136](https://github.com/apache/incubator-tvm/pull/5136) | |
| - Return empty CSourceModule when no lowered_funcs exists in Relay mod [#4847](https://github.com/apache/incubator-tvm/pull/4847) | |
| - Duplicate likely nodes added when loop axis split unevenly [#5084](https://github.com/apache/incubator-tvm/pull/5084) | |
| - Fix incorrect name of calibration mode [#5150](https://github.com/apache/incubator-tvm/pull/5150) | |
| - Remove contrib spatial pack schedule of depthwise convolution [#5148](https://github.com/apache/incubator-tvm/pull/5148) | |
| - Handle empty LLVMModule in GetFunction [#5146](https://github.com/apache/incubator-tvm/pull/5146) | |
| - TVM release process [#5151](https://github.com/apache/incubator-tvm/pull/5151) | |
| - Fix annotate pass static variable [#5023](https://github.com/apache/incubator-tvm/pull/5023) | |
| - Adding support for QNN subtract op [#5153](https://github.com/apache/incubator-tvm/pull/5153) | |
| - Fixed ConvTranspose2D parsing [#5157](https://github.com/apache/incubator-tvm/pull/5157) | |
| - Fix VTA build in CI [#5165](https://github.com/apache/incubator-tvm/pull/5165) | |
| - Create a new parameter --cache-from in tvm/docker/build.sh [#5173](https://github.com/apache/incubator-tvm/pull/5173) | |
| - : nullptr check [#5176](https://github.com/apache/incubator-tvm/pull/5176) | |
| - rocm: fix miopen convolutions [#5179](https://github.com/apache/incubator-tvm/pull/5179) | |
| - Various CI fixes for the VTA and Relay [#5181](https://github.com/apache/incubator-tvm/pull/5181) | |
| - reverse-mode autodiff without any optimization [#5121](https://github.com/apache/incubator-tvm/pull/5121) | |
| - rocm: fix dense_rocblas in strategy, topi [#5191](https://github.com/apache/incubator-tvm/pull/5191) | |
| - Activations for pytorch [#5194](https://github.com/apache/incubator-tvm/pull/5194) | |
| - ref_counter -> ref_counter_ [#5184](https://github.com/apache/incubator-tvm/pull/5184) | |
| - Add warning about nnpack installing googletest [#5185](https://github.com/apache/incubator-tvm/pull/5185) | |
| - Missing vectorize for depthwise conv2d. [#5196](https://github.com/apache/incubator-tvm/pull/5196) | |
| - bugfix in tensorflow space_to_batch_nd [#5175](https://github.com/apache/incubator-tvm/pull/5175) | |
| - Dropouts And InstanceNorm support added [#5203](https://github.com/apache/incubator-tvm/pull/5203) | |
| ## Apr 2020 | |
| ### IR | |
| * [TIR] Enhance Substitute, python bindings for Substitute/PostOrderVisit ([#5400](https://github.com/apache/incubator-tvm/pull/5400)) | |
| * [TIR] Remove ProducerConsumer and AllocateNode::new_expr ([#5333](https://github.com/apache/incubator-tvm/pull/5333)) | |
| * [IR][TRANSFORM] Enable CopyOnWrite for TIR passes. ([#5309](https://github.com/apache/incubator-tvm/pull/5309)) | |
| * [NODE][IR] Introduce StructuralHash for the Unified IR. ([#5160](https://github.com/apache/incubator-tvm/pull/5160)) | |
| * [TIR] Introduce BufferLoad/Store ([#5205](https://github.com/apache/incubator-tvm/pull/5205)) | |
| * [NODE] General serialzation of leaf objects into bytes. ([#5299](https://github.com/apache/incubator-tvm/pull/5299)) | |
| * [POC][IR] Initial stab at std::string->String upgrade ([#5438](https://github.com/apache/incubator-tvm/pull/5438)) | |
| * [NODE][IR] Introduce StructuralEqual Infra for the unified IR. ([#5154](https://github.com/apache/incubator-tvm/pull/5154)) | |
| * [TIR] Make lower_warp_memory support extent(threadIdx.x) < warp_size ([#5307](https://github.com/apache/incubator-tvm/pull/5307)) | |
| * [TE] Support mixing normal and cross-thread reduction ([#5193](https://github.com/apache/incubator-tvm/pull/5193)) | |
| * [TIR][PASS] dtype rewrite for indexing variables ([#5092](https://github.com/apache/incubator-tvm/pull/5092)) | |
| ### Arithmetic | |
| * [Arith] linear system and equation solver ([#5171](https://github.com/apache/incubator-tvm/pull/5171)) | |
| * Improve IntervalSet's floormod ([#5367](https://github.com/apache/incubator-tvm/pull/5367)) | |
| ### Relay | |
| * [BYOC] Bind constant tuples in graph partitioner ([#5476](https://github.com/apache/incubator-tvm/pull/5476)) | |
| * [RELAY][BYOC] Add support for composite functions in BYOC ([#5261](https://github.com/apache/incubator-tvm/pull/5261)) | |
| * [RELAY][BYOC] Register pattern tables from external codegens ([#5262](https://github.com/apache/incubator-tvm/pull/5262)) | |
| * [BYOC] Enhance partitioning and external codegen ([#5310](https://github.com/apache/incubator-tvm/pull/5310)) | |
| * [Relay][ADT]Static Tensor Array ([#5103](https://github.com/apache/incubator-tvm/pull/5103)) | |
| * [BYOC] Refine AnnotateTarget and MergeCompilerRegion Passes ([#5277](https://github.com/apache/incubator-tvm/pull/5277)) | |
| * [BYOC] Use Non-Recursive Visitor/Mutator ([#5410](https://github.com/apache/incubator-tvm/pull/5410)) | |
| * [BYOC] Refine DNNL Codegen ([#5288](https://github.com/apache/incubator-tvm/pull/5288)) | |
| * [RELAY] Non-recursive Graph Vistor and Rewriter ([#4886](https://github.com/apache/incubator-tvm/pull/4886)) | |
| * [Blocksparse] Pipeline for lowering dense model to sparse-dense ([#5377](https://github.com/apache/incubator-tvm/pull/5377)) | |
| * [BYOC] Prevent duplicate outputs in subgraph Tuple ([#5320](https://github.com/apache/incubator-tvm/pull/5320)) | |
| * [RELAY] Re-wrote the Graph Partitioner to support multiple outputs ([#5143](https://github.com/apache/incubator-tvm/pull/5143)) | |
| ### Framework Support | |
| * [Frontend] Asymmetric padding of convolution support ([#4803](https://github.com/apache/incubator-tvm/pull/4803)) | |
| * [ONNX]Pool3d & upsample3d op support ([#5135](https://github.com/apache/incubator-tvm/pull/5135)) | |
| * Add TopK to ONNX Frontend ([#5441](https://github.com/apache/incubator-tvm/pull/5441)) | |
| * Add RoiAlign to Onnx frontend ([#5454](https://github.com/apache/incubator-tvm/pull/5454)) | |
| * [PYTORCH]AvgPool3d, MaxPool3d and Squeeze op support ([#5220](https://github.com/apache/incubator-tvm/pull/5220)) | |
| * [PYTORCH]celu, gelu, selu activations ([#5263](https://github.com/apache/incubator-tvm/pull/5263)) | |
| * [Pytorch]layernorm bug fix and testcase updated ([#5257](https://github.com/apache/incubator-tvm/pull/5257)) | |
| * [PYTORCH]LayerNorm support added ([#5249](https://github.com/apache/incubator-tvm/pull/5249)) | |
| * [RELAY-OP][PYTORCH]GroupNorm op support added ([#5358](https://github.com/apache/incubator-tvm/pull/5358)) | |
| * [TOPI][PYTORCH]Logical & Bitwise operator support ([#5341](https://github.com/apache/incubator-tvm/pull/5341)) | |
| * [PYTORCH]Tensor creation ops support ([#5347](https://github.com/apache/incubator-tvm/pull/5347)) | |
| * [RELAY][PYTORCH]cosh,sinh,log2,log10,log1p op support ([#5395](https://github.com/apache/incubator-tvm/pull/5395)) | |
| * [PYTORCH]Rsub, Embedded, OneHot ops support ([#5434](https://github.com/apache/incubator-tvm/pull/5434)) | |
| * [PYTORCH]Abs, Arange, Softplus ops ([#5295](https://github.com/apache/incubator-tvm/pull/5295)) | |
| * [RELAY][PYTORCH]isNan, isinf, isfinite, ceil, clamp, round ops ([#5316](https://github.com/apache/incubator-tvm/pull/5316)) | |
| * [PYTORCH]Activations for pytorch ([#5194](https://github.com/apache/incubator-tvm/pull/5194)) | |
| * [PYTORCH]Repeat, Reciprocal & Reshape Op support ([#5280](https://github.com/apache/incubator-tvm/pull/5280)) | |
| * [PYTORCH]Reduce_ops support added ([#5308](https://github.com/apache/incubator-tvm/pull/5308)) | |
| * [PYTORCH]Take, Topk op support ([#5332](https://github.com/apache/incubator-tvm/pull/5332)) | |
| * [PYTORCH]Dropouts And InstanceNorm support added ([#5203](https://github.com/apache/incubator-tvm/pull/5203)) | |
| * [PYTORCH]Unary Ops frontend support. ([#5378](https://github.com/apache/incubator-tvm/pull/5378)) | |
| * [PYTORCH]where, addcdiv, addcmul op support ([#5383](https://github.com/apache/incubator-tvm/pull/5383)) | |
| * [Torch] Support Python list, more realistic recurrent networks ([#5306](https://github.com/apache/incubator-tvm/pull/5306)) | |
| * [Torch] Add support for split ([#5174](https://github.com/apache/incubator-tvm/pull/5174)) | |
| * [Frontend][Torch] Fix up graph input handling ([#5204](https://github.com/apache/incubator-tvm/pull/5204)) | |
| * [FRONTEND][TFLITE]Logical not op support ([#5475](https://github.com/apache/incubator-tvm/pull/5475)) | |
| * [TFLITE]Hard Swish & MobilnetV3 model testing ([#5239](https://github.com/apache/incubator-tvm/pull/5239)) | |
| * [FRONTEND][TFLITE]Gather, StridedSlice op support added ([#4788](https://github.com/apache/incubator-tvm/pull/4788)) | |
| * [TFLITE] Match TFLite shape for SSD custom op ([#5473](https://github.com/apache/incubator-tvm/pull/5473)) | |
| * Factor out import of common tflite.Operator in tflite frontend. ([#5355](https://github.com/apache/incubator-tvm/pull/5355)) | |
| * [Frontend][TFLite] support for FILL and SPLIT_V operators ([#5330](https://github.com/apache/incubator-tvm/pull/5330)) | |
| * [Frontend][TFLite] L2_POOL_2D operator ([#5452](https://github.com/apache/incubator-tvm/pull/5452)) | |
| * [TFLite] Add config option to specify FlatBuffers location ([#5425](https://github.com/apache/incubator-tvm/pull/5425)) | |
| * [FRONTEND][TFLITE]Logical not op support ([#5475](https://github.com/apache/incubator-tvm/pull/5475)) | |
| * [TENSORFLOW]reduce ops updated ([#5180](https://github.com/apache/incubator-tvm/pull/5180)) | |
| * [FRONTEND][TENSORFLOW] Fix gather_nd indices ([#5279](https://github.com/apache/incubator-tvm/pull/5279)) | |
| * [Frontend][TensorFlow]Improve TensorFlow Static Shape Tensor Array ([#5243](https://github.com/apache/incubator-tvm/pull/5243)) | |
| * [KERAS]Minimum & AlphaDropout op support ([#5380](https://github.com/apache/incubator-tvm/pull/5380)) | |
| * [KERAS]Upsample3d & ZeroPadding3d op ([#5125](https://github.com/apache/incubator-tvm/pull/5125)) | |
| * [KERAS]Embedding layer ([#5444](https://github.com/apache/incubator-tvm/pull/5444)) | |
| * [FRONTEND][KERAS]Max_pool3d and Averagepool3d operator support ([#5085](https://github.com/apache/incubator-tvm/pull/5085)) | |
| * [RELAY][FRONTEND][CAFFE2] add Mul and ConvTranspose operator ([#5302](https://github.com/apache/incubator-tvm/pull/5302)) | |
| * [MXNET]DepthToSpace & SpaceToDepth Operator ([#5408](https://github.com/apache/incubator-tvm/pull/5408)) | |
| * [MXNET]broadcast and logical op support ([#5461](https://github.com/apache/incubator-tvm/pull/5461)) | |
| * [FRONTEND][MXNET] Use leaky by default for LeakyReLU ([#5192](https://github.com/apache/incubator-tvm/pull/5192)) | |
| * [FRONTEND][MXNET] support elemwise logic ops ([#5361](https://github.com/apache/incubator-tvm/pull/5361)) | |
| * [Frontend|MXNet] SwapAxis operator support ([#5246](https://github.com/apache/incubator-tvm/pull/5246)) | |
| ### Object and Python Frontend | |
| * [PYTHON] Enhance with_attr API, cleanup MakeAPILegacy in testcases ([#5335](https://github.com/apache/incubator-tvm/pull/5335)) | |
| * [Runtime][Object] expose runtime::String to Python ([#5212](https://github.com/apache/incubator-tvm/pull/5212)) | |
| * [PYTHON] Make IntImm more like an integer ([#5232](https://github.com/apache/incubator-tvm/pull/5232)) | |
| * [PY][FFI] Refactor runtime.String to subclass str ([#5426](https://github.com/apache/incubator-tvm/pull/5426)) | |
| * [RUNTIME][IR] Allow non-nullable ObjectRef, introduce Optional<T>. ([#5314](https://github.com/apache/incubator-tvm/pull/5314)) | |
| * [RUNTIME][OBJECT] Introduce static slots for common objects. ([#5423](https://github.com/apache/incubator-tvm/pull/5423)) | |
| * [RUNTIME] Introduce RValue reference(move) support to TypedPackedFunc ([#5271](https://github.com/apache/incubator-tvm/pull/5271)) | |
| * [RUNTIME] Auto conversion from str to runtime::String in PackedFUnc ([#5251](https://github.com/apache/incubator-tvm/pull/5251)) | |
| * [RUNTIME] Improved Packed FFI for optional. ([#5478](https://github.com/apache/incubator-tvm/pull/5478)) | |
| ### Operator support | |
| * [TOPI] Using x86 schedules for ARM conv2d ([#5334](https://github.com/apache/incubator-tvm/pull/5334)) | |
| * [TOPI-ARM] Do not alter layout if layout is NHWC ([#5350](https://github.com/apache/incubator-tvm/pull/5350)) | |
| * [TOPI] Setting workload correctly for Depthwise Spatial conv ARM. ([#5182](https://github.com/apache/incubator-tvm/pull/5182)) | |
| * [Relay][OP] Add fast_erf implementation ([#5241](https://github.com/apache/incubator-tvm/pull/5241)) | |
| * [Topi] Tensorcore support for Conv3D ([#5284](https://github.com/apache/incubator-tvm/pull/5284)) | |
| * [intrin] a few more math functions ([#5468](https://github.com/apache/incubator-tvm/pull/5468)) | |
| * [Intrinsic] Add log1p, ldexp, atan2, hypot, nextafter, copysign ([#5312](https://github.com/apache/incubator-tvm/pull/5312)) | |
| * [relay][topi] Add operation relay.nn.dilate() which calls topi.nn.dilate() ([#5331](https://github.com/apache/incubator-tvm/pull/5331)) | |
| ### Performance and AutoTVM | |
| * [Topi x86] Missing vectorize for depthwise conv2d. ([#5196](https://github.com/apache/incubator-tvm/pull/5196)) | |
| * [TOPI x86] Adding unroll_kw config option for depthwise conv2d. ([#5197](https://github.com/apache/incubator-tvm/pull/5197)) | |
| * [Runtime][Contrib] Support cudnn softmax ([#5214](https://github.com/apache/incubator-tvm/pull/5214)) | |
| * [cuDNN] Add cuDNN grouped convolution support ([#5319](https://github.com/apache/incubator-tvm/pull/5319)) | |
| * [Relay][Topi][AutoTVM] Winograd support for Conv3D ([#5186](https://github.com/apache/incubator-tvm/pull/5186)) | |
| * [TOPI] Improve get_valid_count and nms performance for CUDA ([#5339](https://github.com/apache/incubator-tvm/pull/5339)) | |
| * [Topi][Cuda]Optimizations of global_ave_pool for NHWC layout ([#5450](https://github.com/apache/incubator-tvm/pull/5450)) | |
| ### Backend | |
| * [LLVM] Do not use x86_vcvtph2ps_256 intrinsic with LLVM 11+ ([#5267](https://github.com/apache/incubator-tvm/pull/5267)) | |
| * [LLVM] Use llvm::ElementCount with LLVM 11+ when creating vectors ([#5265](https://github.com/apache/incubator-tvm/pull/5265)) | |
| * [LLVM] Use llvm::FunctionCallee in IRBuilder::CreateCall with LLVM 11+ ([#5338](https://github.com/apache/incubator-tvm/pull/5338)) | |
| * [LLVM] Include Support/Host.h for declaration of getDefaultTargetTriple ([#5268](https://github.com/apache/incubator-tvm/pull/5268)) | |
| * [LLVM] Replace calls to Type::getVectorNumElements ([#5398](https://github.com/apache/incubator-tvm/pull/5398)) | |
| * [LLVM] Use ArrayRef<int> in calls to CreateShuffleVector ([#5399](https://github.com/apache/incubator-tvm/pull/5399)) | |
| * [LLVM] Use llvm::Align with LLVM 11+ to avoid warnings ([#5264](https://github.com/apache/incubator-tvm/pull/5264)) | |
| ### Runtime | |
| * [uTVM][Runtime] Introduce Virtual Memory Allocator to CRT ([#5124](https://github.com/apache/incubator-tvm/pull/5124)) | |
| * [RUNTIME] Initial implementation of Hexagon runtime support ([#5252](https://github.com/apache/incubator-tvm/pull/5252)) | |
| * [Hexagon] Add hexagon_posix.cc to TVM/RT sources in the right place ([#5346](https://github.com/apache/incubator-tvm/pull/5346)) | |
| * [RUNTIME] FastRPC interface for Hexagon runtime ([#5353](https://github.com/apache/incubator-tvm/pull/5353)) | |
| * [RUNTIME][CONTRIB] CoreML Runtime ([#5283](https://github.com/apache/incubator-tvm/pull/5283)) | |
| * [Runtime][Relay][Cleanup] Clean up for memory pass to enable heterogenous execution support. ([#5324](https://github.com/apache/incubator-tvm/pull/5324)) | |
| * [RUNTIME][uTVM] AutoTVM + uTVM for Cortex-M7 ([#5417](https://github.com/apache/incubator-tvm/pull/5417)) | |
| * Windows Support for cpp_rpc ([#4857](https://github.com/apache/incubator-tvm/pull/4857)) | |
| * [RUNTIME] Implement TVMDSOOp(TensorFlow custom op) for TVM runtime ([#4459](https://github.com/apache/incubator-tvm/pull/4459)) | |
| ### QNN and quantization | |
| * [Requantize] Cleanup and Optimize Lowering ([#5286](https://github.com/apache/incubator-tvm/pull/5286)) | |
| * [Topi, ARM] Disbale Winograd for quantized tensors. ([#5363](https://github.com/apache/incubator-tvm/pull/5363)) | |
| * Adding support for TFLite QnnSubtract operator. ([#5230](https://github.com/apache/incubator-tvm/pull/5230)) | |
| * Remove developer facing api from frontend exports. ([#5375](https://github.com/apache/incubator-tvm/pull/5375)) | |
| ### Infra and Refactor | |
| * [REFACTOR][TIR] Migrate LowerTVMBuiltin, InferFragment, LowerThreadAllreduce, ThreadSync to Pass Manager ([#5213](https://github.com/apache/incubator-tvm/pull/5213)) | |
| * [TIR][REFACTOR] Remove te::Tensor dependencies from TIR passes. ([#5372](https://github.com/apache/incubator-tvm/pull/5372)) | |
| * [REFACTOR][TE] Inline -> te/schedule/operation_inline.h ([#5386](https://github.com/apache/incubator-tvm/pull/5386)) | |
| * [TIR] Refactor MakePackedAPI to target dependent stage. ([#5326](https://github.com/apache/incubator-tvm/pull/5326)) | |
| * [REFACTOR] tvm.hybrid -> te.hybrid ([#5223](https://github.com/apache/incubator-tvm/pull/5223)) | |
| * [REFACTOR][TIR] Migrate most of low-level build to use the Pass Manager. ([#5225](https://github.com/apache/incubator-tvm/pull/5225)) | |
| * [TIR][REFACTOR] Migrate low-level passes in tvm.lower to the Pass Manager ([#5364](https://github.com/apache/incubator-tvm/pull/5364)) | |
| * [TIR] Migrate VTA TIR passes to the new pass manager. ([#5397](https://github.com/apache/incubator-tvm/pull/5397)) | |
| * [REFACTOR][TIR] Migrate all low-level passes to the Pass Manager. ([#5233](https://github.com/apache/incubator-tvm/pull/5233)) | |
| * [PY][FFI] Refactor runtime.String to subclass str ([#5426](https://github.com/apache/incubator-tvm/pull/5426)) | |
| * [REFACTOR][TIR] Introduce ExprDeepEqual, Remove IRDeepCompare ([#5206](https://github.com/apache/incubator-tvm/pull/5206)) | |
| * [RELAY] Remove re-exports of tvm.transform ([#5337](https://github.com/apache/incubator-tvm/pull/5337)) | |
| * [ARITH] Remove legacy const pattern functions ([#5387](https://github.com/apache/incubator-tvm/pull/5387)) | |
| * [REFACTOR][ARITH] Remove the legacy Simplify, migrate to Analyzer. ([#5385](https://github.com/apache/incubator-tvm/pull/5385)) | |
| * [TIR][REFACTOR] RewriteForTensorCore -> te/schedule ([#5379](https://github.com/apache/incubator-tvm/pull/5379)) | |
| * [RELAY] Move frontend utils ([#5345](https://github.com/apache/incubator-tvm/pull/5345)) | |
| * [REFACTOR][IR] Move to runtime::String ([#5276](https://github.com/apache/incubator-tvm/pull/5276)) | |
| * [REFACTOR][IR] kExternalSymbol -> kGlobalSymbol ([#5211](https://github.com/apache/incubator-tvm/pull/5211)) | |
| * [REFACTOR][IR] Remove PrimExpr from String ([#5311](https://github.com/apache/incubator-tvm/pull/5311)) | |
| * [Topi] Breakdown topi.cc into smaller files ([#5253](https://github.com/apache/incubator-tvm/pull/5253)) | |
| * [Refactor] Add memoized expr translator for use by backend codegen ([#5325](https://github.com/apache/incubator-tvm/pull/5325)) | |
| * [CodeGen] Cleanup generated code ([#5424](https://github.com/apache/incubator-tvm/pull/5424)) | |
| * [IR][Debug] Add dump and print for debugging (NFC) ([#5207](https://github.com/apache/incubator-tvm/pull/5207)) | |
| * Customize SI prefix in logging ([#5411](https://github.com/apache/incubator-tvm/pull/5411)) | |
| * [TIR][REFACTOR] Remove ir_pass in favor of analysis/transform. ([#5415](https://github.com/apache/incubator-tvm/pull/5415)) | |
| * Legalize - Use Non-recursive Rewriter. ([#5296](https://github.com/apache/incubator-tvm/pull/5296)) | |
| ### CI and tests | |
| * [CI] Fix the hexagon string ([#5304](https://github.com/apache/incubator-tvm/pull/5304)) | |
| * [CI] Temporary disable CRT test ([#5297](https://github.com/apache/incubator-tvm/pull/5297)) | |
| * [CI][DOCKER] Update ci-gpu to the lastest ([#5469](https://github.com/apache/incubator-tvm/pull/5469)) | |
| * Removing older Object detection TFlite test ([#5477](https://github.com/apache/incubator-tvm/pull/5477)) | |
| * [CI] Enable tsim and fsim for GPU build to avoid pack_lib error ([#5352](https://github.com/apache/incubator-tvm/pull/5352)) | |
| * [CI] Update MxNet to 1.6.0 with MKL ([#5240](https://github.com/apache/incubator-tvm/pull/5240)) | |
| * [LINT] Remove scalalint from lint deps ([#5269](https://github.com/apache/incubator-tvm/pull/5269)) | |
| * [Rust][CI] Restore Rust CI ([#5137](https://github.com/apache/incubator-tvm/pull/5137)) | |
| * [RFC] Pytest environment improvements ([#5421](https://github.com/apache/incubator-tvm/pull/5421)) | |
| * [CI] Migrate Tensorflow and Tensorflow lite in CI to 2.1.0 ([#5392](https://github.com/apache/incubator-tvm/pull/5392)) | |
| * [CI] Fix build.sh to propagate --network=host to the docker build command ([#5336](https://github.com/apache/incubator-tvm/pull/5336)) | |
| * [CI] Include local Docker images as source for layers ([#5466](https://github.com/apache/incubator-tvm/pull/5466)) | |
| * [TFLite Runtime] Add TFLite Runtime dependencies to CI CPU docker build ([#5437](https://github.com/apache/incubator-tvm/pull/5437)) | |
| ### Docs | |
| * [DOCS] Migrate some markdowns to rst, fix sphinx3 warnings ([#5416](https://github.com/apache/incubator-tvm/pull/5416)) | |
| * [DOCS] Misc docs improvements ([#5222](https://github.com/apache/incubator-tvm/pull/5222)) | |
| * [DOCS] Bring relay docs to the top-level flat view ([#5343](https://github.com/apache/incubator-tvm/pull/5343)) | |
| * [DOCS] Reduce artifcats generated by sphinx gallery ([#5208](https://github.com/apache/incubator-tvm/pull/5208)) | |
| * [DOCS] Use https link ([#5183](https://github.com/apache/incubator-tvm/pull/5183)) | |
| * [DOCSTRING]missing function parameters updated ([#5228](https://github.com/apache/incubator-tvm/pull/5228)) | |
| * [DOCS] Migrate HLS documents from md to rst ([#5419](https://github.com/apache/incubator-tvm/pull/5419)) | |
| * [BYOC] Add example of Composite + Annotate for DNNL fused op ([#5272](https://github.com/apache/incubator-tvm/pull/5272)) | |
| * [Tutorial, QNN] Add tutorial for loading quantized PyTorch model ([#5321](https://github.com/apache/incubator-tvm/pull/5321)) | |
| * [Docs] VTA install doc migration from md to rst ([#5442](https://github.com/apache/incubator-tvm/pull/5442)) | |
| * [TVM][docs] compiler version in docs ([#5281](https://github.com/apache/incubator-tvm/pull/5281)) | |
| ### Fixes | |
| * [BUGFIX] Fix CRT static test bug ([#5293](https://github.com/apache/incubator-tvm/pull/5293)) | |
| * [RUNTIME] Quick fix PackedFunc String passing ([#5266](https://github.com/apache/incubator-tvm/pull/5266)) | |
| * [TIR] Fix perf regression of tir refactor ([#5258](https://github.com/apache/incubator-tvm/pull/5258)) | |
| * [BUGFIX]bugfix in tensorflow space_to_batch_nd ([#5175](https://github.com/apache/incubator-tvm/pull/5175)) | |
| * [RUNTIME][CRT]Compilation warnings fixed for 32bit and 64bit compilation ([#5349](https://github.com/apache/incubator-tvm/pull/5349)) | |
| * [BYOC][FIX] Fix typo in "default" ([#5348](https://github.com/apache/incubator-tvm/pull/5348)) | |
| * [RELAY][FIX] Fix hang in MergeCompilerRegions ([#5227](https://github.com/apache/incubator-tvm/pull/5227)) | |
| * [RELAY] Fixes to MergeCompilerRegions ([#5195](https://github.com/apache/incubator-tvm/pull/5195)) | |
| * [LLVM] Fix generation of LLVM intrinsics ([#5282](https://github.com/apache/incubator-tvm/pull/5282)) | |
| * Fix setting up hints for getaddrinfo ([#2872](https://github.com/apache/incubator-tvm/pull/2872)) | |
| * [Fix] Add ConstantNode to IsAtomic ([#5457](https://github.com/apache/incubator-tvm/pull/5457)) | |
| * [BUGFIX][IR] Fix String SEqual ([#5275](https://github.com/apache/incubator-tvm/pull/5275)) | |
| * [FIX][VM] fix fuse over functions that are handled by external codegen ([#5365](https://github.com/apache/incubator-tvm/pull/5365)) | |
| * [Relay] Fix memory leak when accessing NDArray ([#5413](https://github.com/apache/incubator-tvm/pull/5413)) | |
| * [Fix] Remove the duplicate PrintIR pass in Relay ([#5403](https://github.com/apache/incubator-tvm/pull/5403)) | |
| * [TIR] Fix lower_warp_memory ([#5247](https://github.com/apache/incubator-tvm/pull/5247)) | |
| * [TIR] Fix lower_warp_memory when there are >1 warp buffers ([#5368](https://github.com/apache/incubator-tvm/pull/5368)) | |
| * [External codegen] Add test cases for fused ops with manual annotation ([#4741](https://github.com/apache/incubator-tvm/pull/4741)) | |
| * Fix intel conv2d auto tune ([#5200](https://github.com/apache/incubator-tvm/pull/5200)) | |
| * [CodeGen][CUDA] Fix bugs ([#5209](https://github.com/apache/incubator-tvm/pull/5209)) | |
| * Don't remove() TemporaryFile in __del__. ([#5414](https://github.com/apache/incubator-tvm/pull/5414)) | |
| * Fix test_ir_type. ([#5390](https://github.com/apache/incubator-tvm/pull/5390)) | |
| * [Relay][Frontend][Onnx] Fix multiple identical inputs bug ([#5389](https://github.com/apache/incubator-tvm/pull/5389)) | |
| * [Relay][Strategy] Add cuda target check to dense tensorcore schedule. ([#5376](https://github.com/apache/incubator-tvm/pull/5376)) | |
| * Tf2 test fixups ([#5391](https://github.com/apache/incubator-tvm/pull/5391)) | |
| * [REALY][OP] fix typo ([#5315](https://github.com/apache/incubator-tvm/pull/5315)) | |
| * [Node] Provide guide to user who has difficulty register SEqualReduce ([#5300](https://github.com/apache/incubator-tvm/pull/5300)) | |
| * [NDArray] Set NDArray::Container.shape_ in NDArray::FromDLPack ([#5301](https://github.com/apache/incubator-tvm/pull/5301)) | |
| * fix miopen padding ([#5433](https://github.com/apache/incubator-tvm/pull/5433)) | |
| * misc fixes for ROCm ([#5431](https://github.com/apache/incubator-tvm/pull/5431)) | |
| * Create loops according to storage scope and thread hierarchies ([#5190](https://github.com/apache/incubator-tvm/pull/5190)) | |
| * [RELAY] Partition graph codestyle fixes ([#5202](https://github.com/apache/incubator-tvm/pull/5202)) | |
| * [TE][BuildModule] Fix import in dump pass ir ([#5327](https://github.com/apache/incubator-tvm/pull/5327)) | |
| * docker: Drop caffe2 download progess bars ([#5359](https://github.com/apache/incubator-tvm/pull/5359)) | |
| * [Fix][VM] Fix copy constructor ([#5237](https://github.com/apache/incubator-tvm/pull/5237)) | |
| * [Relay][Tutorial][Fix] Fixed typo and type mismatch in relay infrastructure tutorial ([#5259](https://github.com/apache/incubator-tvm/pull/5259)) | |
| * Corrected TVM autotuning on GPU ([#5432](https://github.com/apache/incubator-tvm/pull/5432)) | |
| * [CODEGEN][CUDA] Fix vector load ([#5226](https://github.com/apache/incubator-tvm/pull/5226)) | |
| * [Fontend][Pytorch] Fix translation of transpose when axis argument is as a list ([#5451](https://github.com/apache/incubator-tvm/pull/5451)) | |
| * [TE] Minor bugfix in message_passing.cc ([#5254](https://github.com/apache/incubator-tvm/pull/5254)) | |
| * [CODEGEN][CUDA] Fix a bug when vectorized load&store was involved for… ([#5428](https://github.com/apache/incubator-tvm/pull/5428)) | |
| * fix to skip node not in graph. ([#5238](https://github.com/apache/incubator-tvm/pull/5238)) | |
| * fix [#5388](https://github.com/apache/incubator-tvm/pull/5388) [RUNTIME][VULKAN] vkBuffer released before memory copy command se… ([#5418](https://github.com/apache/incubator-tvm/pull/5418)) | |
| * [BUGFIX][RELAY]fix a minor error in device_annotation ([#5291](https://github.com/apache/incubator-tvm/pull/5291)) | |
| * [VTA] Fix VTA compile issue ([#5481](https://github.com/apache/incubator-tvm/pull/5481)) | |
| * [RUNTIME][CRT] scalar's ndim is 0 ([#5344](https://github.com/apache/incubator-tvm/pull/5344)) | |
| ### Submodule | |
| * [SUBMODULE] Update dmlc-core to latest ([#5401](https://github.com/apache/incubator-tvm/pull/5401)) | |
| ## May 2020 | |
| ### BYOC (Bring Your Own Codegen) | |
| - Add additional check before re-using the cached match [#5552](https://github.com/apache/incubator-tvm/pull/5552) | |
| - Remove kCompiler attr from external functions [#5615](https://github.com/apache/incubator-tvm/pull/5615) | |
| - Pattern Language MergeComposite [#5656](https://github.com/apache/incubator-tvm/pull/5656) | |
| - Support Tuple Output in C/DNNL Codegen [#5701](https://github.com/apache/incubator-tvm/pull/5701) | |
| - Infer types in MergeComposite [#5766](https://github.com/apache/incubator-tvm/pull/5766) | |
| ### Pattern Language | |
| - Convert PatternGrouper to do pre-order, non-recursive analysis [#5653](https://github.com/apache/incubator-tvm/pull/5653) | |
| - Remove constants from partitioned functions [#5663](https://github.com/apache/incubator-tvm/pull/5663) | |
| - Add a check for null function attributes [#5674](https://github.com/apache/incubator-tvm/pull/5674) | |
| - Add ConstantPattern [#5689](https://github.com/apache/incubator-tvm/pull/5689) | |
| - Conditionally Embedding Constants in Partitioned Functions [#5693](https://github.com/apache/incubator-tvm/pull/5693) | |
| - Simplify Pattern API Implementations [#5703](https://github.com/apache/incubator-tvm/pull/5703) | |
| - Add ShapePattern and DataTypePattern [#5760](https://github.com/apache/incubator-tvm/pull/5760) | |
| - Remove unnecessary print [#5642](https://github.com/apache/incubator-tvm/pull/5642) | |
| ### AutoTVM | |
| - Fix the runtime raise error [#5586](https://github.com/apache/incubator-tvm/pull/5586) | |
| - Update XGBoost verbosity option [#5649](https://github.com/apache/incubator-tvm/pull/5649) | |
| ### TIR | |
| - text format printer considering future parsing use [#5483](https://github.com/apache/incubator-tvm/pull/5483) | |
| - Remove buffer params from pass config. [#5652](https://github.com/apache/incubator-tvm/pull/5652) | |
| ### Relay | |
| - ReduceLogSumExp Operator support [#5453](https://github.com/apache/incubator-tvm/pull/5453) | |
| - Math ops added [#5502](https://github.com/apache/incubator-tvm/pull/5502) | |
| - enable blocking format in x86 conv2d and fold scale axis [#5357](https://github.com/apache/incubator-tvm/pull/5357) | |
| - Fixed bug in attribute parsing for pool layers. [#5582](https://github.com/apache/incubator-tvm/pull/5582) | |
| - Support symbolic newshape for Reshape [#5429](https://github.com/apache/incubator-tvm/pull/5429) | |
| - Specify additional layouts in convert layout pass [#5422](https://github.com/apache/incubator-tvm/pull/5422) | |
| - Safe check added for Merge Composite Call Node [#5562](https://github.com/apache/incubator-tvm/pull/5562) | |
| - Memory planner (part 1) [#5144](https://github.com/apache/incubator-tvm/pull/5144) | |
| - Improve Shape Func handling for Tuple inputs [#5467](https://github.com/apache/incubator-tvm/pull/5467) | |
| - Relay updated with String [#5578](https://github.com/apache/incubator-tvm/pull/5578) | |
| - Fix the creation of tuple of tuples in PartitionGraph [#5616](https://github.com/apache/incubator-tvm/pull/5616) | |
| - Resize3d, Upsample3d op support [#5633](https://github.com/apache/incubator-tvm/pull/5633) | |
| - Preserve type information in Merge Composite [#5640](https://github.com/apache/incubator-tvm/pull/5640) | |
| - Add operator Correlation [#5628](https://github.com/apache/incubator-tvm/pull/5628) | |
| - Move compiler_begin/end_op to local static objects [#5622](https://github.com/apache/incubator-tvm/pull/5622) | |
| - affine_grid and grid_sample [#5657](https://github.com/apache/incubator-tvm/pull/5657) | |
| - Support symbolic TopK, Ones, Zeros and Full [#5459](https://github.com/apache/incubator-tvm/pull/5459) | |
| - Fix dataflow_pattern.rewrite() hang if Match in IR [#5680](https://github.com/apache/incubator-tvm/pull/5680) | |
| - Fix segfault in pretty print when ObjectRef is null [#5681](https://github.com/apache/incubator-tvm/pull/5681) | |
| - move fallback_device to config [#5690](https://github.com/apache/incubator-tvm/pull/5690) | |
| - Replace build_config with PassContext [#5698](https://github.com/apache/incubator-tvm/pull/5698) | |
| - Clear compile engine after task extraction [#5724](https://github.com/apache/incubator-tvm/pull/5724) | |
| - Sparse to dense operator [#5447](https://github.com/apache/incubator-tvm/pull/5447) | |
| - support dynamic NMS(Non Maximum Suppression), symbolic begin, end, and strides for strided_slice [#4312](https://github.com/apache/incubator-tvm/pull/4312) | |
| - Conv3d_transpose op support added [#5737](https://github.com/apache/incubator-tvm/pull/5737) | |
| - Fix for recursive let [#5757](https://github.com/apache/incubator-tvm/pull/5757) | |
| - Add operation gather to relay. [#5716](https://github.com/apache/incubator-tvm/pull/5716) | |
| - Fix Calibration Pass to Support Modules with Multiple Functions [#5768](https://github.com/apache/incubator-tvm/pull/5768) | |
| - Add storage_order ignore in pooling layer. [#5781](https://github.com/apache/incubator-tvm/pull/5781) | |
| ### Topi | |
| - Optimization of Conv2d Winograd algorithm on Tensor … [#5485](https://github.com/apache/incubator-tvm/pull/5485) | |
| - AutoTVM incorrect measurement [#5511](https://github.com/apache/incubator-tvm/pull/5511) | |
| - Fix bifrost spatial packing conv2d auto tune [#5684](https://github.com/apache/incubator-tvm/pull/5684) | |
| - Fix reshape usage in ARM schedule [#5732](https://github.com/apache/incubator-tvm/pull/5732) | |
| - block sparse dense on cuda [#5746](https://github.com/apache/incubator-tvm/pull/5746) | |
| - Improve CUDA softmax scheduling [#5600](https://github.com/apache/incubator-tvm/pull/5600) | |
| - pass-by-value -> pass-by-const-reference [#5783](https://github.com/apache/incubator-tvm/pull/5783) | |
| - fix sparse dense schedule on cuda [#5803](https://github.com/apache/incubator-tvm/pull/5803) | |
| - fix strategy for sparse dense cuda [#5782](https://github.com/apache/incubator-tvm/pull/5782) | |
| ### Arithmetic | |
| - Handle likely in IRMutatorWithAnalyzer [#5665](https://github.com/apache/incubator-tvm/pull/5665) | |
| - ExtendedEuclidean merge impl to int_operator [#5625](https://github.com/apache/incubator-tvm/pull/5625) | |
| - fix a min/max simplify bug [#5749](https://github.com/apache/incubator-tvm/pull/5749) | |
| - fix a min/max simplify bug [#5761](https://github.com/apache/incubator-tvm/pull/5761) | |
| ### Runtime | |
| - Fix workspace [#5503](https://github.com/apache/incubator-tvm/pull/5503) | |
| - Store nullptr PackedFunc as nullptr for better error propagation [#5540](https://github.com/apache/incubator-tvm/pull/5540) | |
| - WebGPU support [#5545](https://github.com/apache/incubator-tvm/pull/5545) | |
| - Hexagon driver for offloading kernels to simulator [#5492](https://github.com/apache/incubator-tvm/pull/5492) | |
| - Setup lint, doc, test [#5556](https://github.com/apache/incubator-tvm/pull/5556) | |
| - TVM WebAssembly JS Runtime [#5506](https://github.com/apache/incubator-tvm/pull/5506) | |
| - Improve PackedFunc robustness [#5517](https://github.com/apache/incubator-tvm/pull/5517) | |
| - Seg fault in WorkspacePool's destructor (#5632) [#5636](https://github.com/apache/incubator-tvm/pull/5636) | |
| - Introduce runtime::Array [#5585](https://github.com/apache/incubator-tvm/pull/5585) | |
| - Resolve constexpr issue in debug mode. [#5651](https://github.com/apache/incubator-tvm/pull/5651) | |
| - Add compile_shared option to linux compile utility fn [#5751](https://github.com/apache/incubator-tvm/pull/5751) | |
| ### RPC | |
| - Call sync in CopyFromRemote and CopyToRemote [#5512](https://github.com/apache/incubator-tvm/pull/5512) | |
| - Fix the multihop cpu case [#5522](https://github.com/apache/incubator-tvm/pull/5522) | |
| - Improve RPCServer AsyncIO support. [#5544](https://github.com/apache/incubator-tvm/pull/5544) | |
| ### ONNX | |
| - LpPool Support added [#5696](https://github.com/apache/incubator-tvm/pull/5696) | |
| - Skip ADD inside Gemm op when vector is zero [#5697](https://github.com/apache/incubator-tvm/pull/5697) | |
| - ReduceL1, ReduceL2, ReduceSumSquare, ReduceLogSum ops added [#5721](https://github.com/apache/incubator-tvm/pull/5721) | |
| - MaxRoiPool, Mod & Xor op support added [#5729](https://github.com/apache/incubator-tvm/pull/5729) | |
| - Skip multiply with 1.0f constant for GEMM import [#5800](https://github.com/apache/incubator-tvm/pull/5800) | |
| ### Tensorflow | |
| - StatefulPartitionedCall/PartitionedCall Ops support added [#5617](https://github.com/apache/incubator-tvm/pull/5617) | |
| - Don't add cast for batch norm when type isn't changing [#5731](https://github.com/apache/incubator-tvm/pull/5731) | |
| - Conv3d Transpose OP added [#5775](https://github.com/apache/incubator-tvm/pull/5775) | |
| ### Pytorch | |
| - expand bug fix [#5576](https://github.com/apache/incubator-tvm/pull/5576) | |
| - Support max_pool2d_with_indices [#5549](https://github.com/apache/incubator-tvm/pull/5549) | |
| - Add prim::device op [#5584](https://github.com/apache/incubator-tvm/pull/5584) | |
| - ImplicitTensorToNum support added [#5603](https://github.com/apache/incubator-tvm/pull/5603) | |
| - Matmul fix for batch_matmul [#5604](https://github.com/apache/incubator-tvm/pull/5604) | |
| - ReflectionPad2d op [#5624](https://github.com/apache/incubator-tvm/pull/5624) | |
| - Padding op support [#5638](https://github.com/apache/incubator-tvm/pull/5638) | |
| - Minor bug fixes [#5683](https://github.com/apache/incubator-tvm/pull/5683) | |
| - floor_divide support for squeezenet [#5702](https://github.com/apache/incubator-tvm/pull/5702) | |
| - ReplicationPad support added [#5708](https://github.com/apache/incubator-tvm/pull/5708) | |
| - aten::norm support added [#5776](https://github.com/apache/incubator-tvm/pull/5776) | |
| ### Mxnet | |
| - broadcast and logical op support [#5461](https://github.com/apache/incubator-tvm/pull/5461) | |
| - MaxPool3d and AvgPool3d Ops support added [#5614](https://github.com/apache/incubator-tvm/pull/5614) | |
| - Softmin, trunc op support added [#5715](https://github.com/apache/incubator-tvm/pull/5715) | |
| - conv3d and conv3d_transpose addedx [#5814](https://github.com/apache/incubator-tvm/pull/5814) | |
| ### Tflite | |
| - Model importer to be compatible with tflite 2.1.0 [#5497](https://github.com/apache/incubator-tvm/pull/5497) | |
| - Nit: Function names made consistent [#5515](https://github.com/apache/incubator-tvm/pull/5515) | |
| - Select op support for tflite frontend [#5486](https://github.com/apache/incubator-tvm/pull/5486) | |
| - GATHER_ND [#5508](https://github.com/apache/incubator-tvm/pull/5508) | |
| - Quantize & Dequantize op [#5394](https://github.com/apache/incubator-tvm/pull/5394) | |
| ### Other frontend | |
| - Fully connected op conversion made in sync with TFLite [#5510](https://github.com/apache/incubator-tvm/pull/5510) | |
| - ADD_N operator [#5474](https://github.com/apache/incubator-tvm/pull/5474) | |
| - onnx, mxnet, pytorch mathops added [#5561](https://github.com/apache/incubator-tvm/pull/5561) | |
| - abs, round, reciprocal, sign, softsign, hard_sigmoid ops support [#5587](https://github.com/apache/incubator-tvm/pull/5587) | |
| - Gather nd bug fix for one dim support in tensorflow [#5588](https://github.com/apache/incubator-tvm/pull/5588) | |
| - Add parser support for shape and range [#5329](https://github.com/apache/incubator-tvm/pull/5329) | |
| - Darknet support batch size for yolo [#5688](https://github.com/apache/incubator-tvm/pull/5688) | |
| - Improve Control Flow and TensorArray [#5699](https://github.com/apache/incubator-tvm/pull/5699) | |
| - Improve TF Parser to keep output nodes for saved_model [#5794](https://github.com/apache/incubator-tvm/pull/5794) | |
| - Add parser support for relu6, leaky_relu, relu_n1_to_1, log_softmax [#4805](https://github.com/apache/incubator-tvm/pull/4805) | |
| ### Docs | |
| - Fix bad restructured text formatting for VTA install guide [#5541](https://github.com/apache/incubator-tvm/pull/5541) | |
| - Improve document in reflection [#5593](https://github.com/apache/incubator-tvm/pull/5593) | |
| - Move the api docs to the api subfolder [#5626](https://github.com/apache/incubator-tvm/pull/5626) | |
| - Fix the QNN TFLite tutorial build [#5641](https://github.com/apache/incubator-tvm/pull/5641) | |
| - Clarify downstream consistency of TVMArgTypeCode [#5742](https://github.com/apache/incubator-tvm/pull/5742) | |
| ### CI | |
| - Install wasmtime for WebAssembly tests [#5494](https://github.com/apache/incubator-tvm/pull/5494) | |
| - Update Jenkins ci-cpu to bionic [#5555](https://github.com/apache/incubator-tvm/pull/5555) | |
| - Update the ci-gpu to the lastest build with the new vulkansdk. [#5571](https://github.com/apache/incubator-tvm/pull/5571) | |
| - Fix clang-format error [#5577](https://github.com/apache/incubator-tvm/pull/5577) | |
| - Enable llvm-11 and llvm-10 in build tests, recover webdocs. [#5579](https://github.com/apache/incubator-tvm/pull/5579) | |
| - Update ci-lint to use the latest image that contains clang-format [#5568](https://github.com/apache/incubator-tvm/pull/5568) | |
| - reintroduce docker stage for wasm tests [#5565](https://github.com/apache/incubator-tvm/pull/5565) | |
| - Allow CI_PYTEST_ADD_OPTIONS to be unbound. [#5644](https://github.com/apache/incubator-tvm/pull/5644) | |
| - Add log check to the sphinx gallery docs [#5643](https://github.com/apache/incubator-tvm/pull/5643) | |
| - Move cpu-only frontend tests to a CPU stage [#5807](https://github.com/apache/incubator-tvm/pull/5807) | |
| - Limit number of threads in all jobs [#5815](https://github.com/apache/incubator-tvm/pull/5815) | |
| ### Refactor | |
| - Non recursive partitioning [#5493](https://github.com/apache/incubator-tvm/pull/5493) | |
| - Modularize the RPC infra [#5484](https://github.com/apache/incubator-tvm/pull/5484) | |
| - IRModule is updated with String [#5523](https://github.com/apache/incubator-tvm/pull/5523) | |
| - IR is updated with String [#5547](https://github.com/apache/incubator-tvm/pull/5547) | |
| - Streamline ir/op Registry [#5609](https://github.com/apache/incubator-tvm/pull/5609) | |
| - Migrate IRModule ObjectRef to not-null [#5654](https://github.com/apache/incubator-tvm/pull/5654) | |
| - Migrate BuildConfig to PassContext. [#5668](https://github.com/apache/incubator-tvm/pull/5668) | |
| - std::string -> String Migration in TIR nodes [#5596](https://github.com/apache/incubator-tvm/pull/5596) | |
| - relay.op.Op -> tvm.ir.Op [#5705](https://github.com/apache/incubator-tvm/pull/5705) | |
| - Separate ArgTypeCode from DLDataTypeCode [#5730](https://github.com/apache/incubator-tvm/pull/5730) | |
| - Remove legacy compute_expr.h [#5738](https://github.com/apache/incubator-tvm/pull/5738) | |
| - Call::Halide => ProducerLoad, DSL/TIR decouple. [#5743](https://github.com/apache/incubator-tvm/pull/5743) | |
| - Provide->ProducerStore, Realize->ProducerRealize. [#5750](https://github.com/apache/incubator-tvm/pull/5750) | |
| - Migrate the tvm/tir/expr.h to constructor [#5773](https://github.com/apache/incubator-tvm/pull/5773) | |
| - Migrate tir/stmt.h to use constructor. [#5778](https://github.com/apache/incubator-tvm/pull/5778) | |
| - Migrate all Object construction to constructor. [#5784](https://github.com/apache/incubator-tvm/pull/5784) | |
| - Cleanup unused classes [#5789](https://github.com/apache/incubator-tvm/pull/5789) | |
| - Finish std::string->String updates [#5793](https://github.com/apache/incubator-tvm/pull/5793) | |
| - Add tir prefix to type keys [#5802](https://github.com/apache/incubator-tvm/pull/5802) | |
| ### Bug Fixes | |
| - Fix bug in rpc ring buffer shrink [#5516](https://github.com/apache/incubator-tvm/pull/5516) | |
| - Fix remote device sync [#5538](https://github.com/apache/incubator-tvm/pull/5538) | |
| - Fix bug in rpc ring buffer shrink (#5516) [#5537](https://github.com/apache/incubator-tvm/pull/5537) | |
| - RPC Server error fix on Pynq FPGA [#5607](https://github.com/apache/incubator-tvm/pull/5607) | |
| - Fix FloorMod Simplifier [#5509](https://github.com/apache/incubator-tvm/pull/5509) | |
| - Fix FloorMod Simplifier [#5505](https://github.com/apache/incubator-tvm/pull/5505) | |
| - Fix Python debugger segfaults with TVM built with LLVM [#5685](https://github.com/apache/incubator-tvm/pull/5685) | |
| - Fix Compilation Error in CRT [#5713](https://github.com/apache/incubator-tvm/pull/5713) | |
| - Fix runtime::String backward compatibility in JSON [#5725](https://github.com/apache/incubator-tvm/pull/5725) | |
| - Allow RPCWrappedFunc to rewrite runtime::String as std::string [#5796](https://github.com/apache/incubator-tvm/pull/5796) | |
| - Fix reshape [#5739](https://github.com/apache/incubator-tvm/pull/5739) | |
| - Make "none" DataType explicit [#5491](https://github.com/apache/incubator-tvm/pull/5491) | |
| - Change "scalar" and "stack" in IDL from "inrout" to "in" [#5487](https://github.com/apache/incubator-tvm/pull/5487) | |
| - Link necessary libraries when building runtime for Android [#5496](https://github.com/apache/incubator-tvm/pull/5496) | |
| - Fixes for wasm32 target [#5489](https://github.com/apache/incubator-tvm/pull/5489) | |
| - Reset target and wait for runtime initialization on connect. [#5499](https://github.com/apache/incubator-tvm/pull/5499) | |
| - bump tophub rocm version [#5504](https://github.com/apache/incubator-tvm/pull/5504) | |
| - Support CallNode inputs in qnn.concatenate [#5360](https://github.com/apache/incubator-tvm/pull/5360) | |
| - Improve commentary for RingBuffer [#5518](https://github.com/apache/incubator-tvm/pull/5518) | |
| - Add unit tests for ONNX PRelu and fix importer to pass them. [#5521](https://github.com/apache/incubator-tvm/pull/5521) | |
| - LRN only supports 4D tensors, remove it from alter_op_layout [#5520](https://github.com/apache/incubator-tvm/pull/5520) | |
| - Fix an issue with ONNX Upsample [#5530](https://github.com/apache/incubator-tvm/pull/5530) | |
| - Cache PrimExpr instead of raw pointers in bound analyzer [#5533](https://github.com/apache/incubator-tvm/pull/5533) | |
| - fix a few bugs with shape inference and types in the ONNX importer [#5534](https://github.com/apache/incubator-tvm/pull/5534) | |
| - FP32 and Quantized Object Detection Model [#5479](https://github.com/apache/incubator-tvm/pull/5479) | |
| - Add Onnx Pad v11 [#5539](https://github.com/apache/incubator-tvm/pull/5539) | |
| - Changes to cpp_rpc to make it work on Android (+ Hexagon offloading) [#5535](https://github.com/apache/incubator-tvm/pull/5535) | |
| - fix to reduce RAM size during loading model [#5507](https://github.com/apache/incubator-tvm/pull/5507) | |
| - Fix MakeLoopNest for warp memory [#5382](https://github.com/apache/incubator-tvm/pull/5382) | |
| - Add first stage of updating and rewriting Rust bindings. [#5526](https://github.com/apache/incubator-tvm/pull/5526) | |
| - Load platform specific lib for tvmdsoop instead of the hard-coded tvm_dso_op.so [#5542](https://github.com/apache/incubator-tvm/pull/5542) | |
| - Add tests for running micro on native arm hardware [#5546](https://github.com/apache/incubator-tvm/pull/5546) | |
| - Apparently, ONNX Conv with no 'pads' defaults to zero padding [#5548](https://github.com/apache/incubator-tvm/pull/5548) | |
| - clang-format the h,cc,m files. [#5557](https://github.com/apache/incubator-tvm/pull/5557) | |
| - Fix conv2d alter op for arm cpu [#5532](https://github.com/apache/incubator-tvm/pull/5532) | |
| - Fix topi test (/topi/tests/python/test_topi_conv2d_nhwc_winograd.py) for non tensorcore CI. [#5563](https://github.com/apache/incubator-tvm/pull/5563) | |
| - Add clang-format and nodejs to ci-lint [#5567](https://github.com/apache/incubator-tvm/pull/5567) | |
| - Enable clang-format. [#5572](https://github.com/apache/incubator-tvm/pull/5572) | |
| - Allow ubuntu_install_darknet.sh to work in both 18.04 and 16.04 [#5574](https://github.com/apache/incubator-tvm/pull/5574) | |
| - Add a quantized conv2 unit test for the tflite front-end [#5558](https://github.com/apache/incubator-tvm/pull/5558) | |
| - Fix JSON graph dumping. [#5591](https://github.com/apache/incubator-tvm/pull/5591) | |
| - Warp level reduction support for CUDA [#5498](https://github.com/apache/incubator-tvm/pull/5498) | |
| - One more fix for concurrency count [#5589](https://github.com/apache/incubator-tvm/pull/5589) | |
| - Improve robustness of the docs build [#5583](https://github.com/apache/incubator-tvm/pull/5583) | |
| - Phase out WebGL [#5570](https://github.com/apache/incubator-tvm/pull/5570) | |
| - Fix vulkansdk in the ci-gpu and upgrade to 1.2.135 [#5566](https://github.com/apache/incubator-tvm/pull/5566) | |
| - Update ci-cpu to bionic [#5554](https://github.com/apache/incubator-tvm/pull/5554) | |
| - Overestimate binary size for microTVM compiled binaries. [#5590](https://github.com/apache/incubator-tvm/pull/5590) | |
| - Fix bug and re-enable RPC execution test [#5436](https://github.com/apache/incubator-tvm/pull/5436) | |
| - Add ostream formatters for TargetPtr/TargetVal. [#5592](https://github.com/apache/incubator-tvm/pull/5592) | |
| - Pattern Language, Matcher, Rewriter, and Function Paritioner [#5231](https://github.com/apache/incubator-tvm/pull/5231) | |
| - Fix cross thread reduction [#5551](https://github.com/apache/incubator-tvm/pull/5551) | |
| - Fix TVMArray layout on device [#5599](https://github.com/apache/incubator-tvm/pull/5599) | |
| - Add debug mode to tempdir() [#5581](https://github.com/apache/incubator-tvm/pull/5581) | |
| - Represent alignment information in LLVM IR [#5598](https://github.com/apache/incubator-tvm/pull/5598) | |
| - Fix codegen for warp shuffle intrinsics [#5606](https://github.com/apache/incubator-tvm/pull/5606) | |
| - Fix Topological Order calculation for DFPattern Language [#5612](https://github.com/apache/incubator-tvm/pull/5612) | |
| - Fix a typo. [#5611](https://github.com/apache/incubator-tvm/pull/5611) | |
| - Global MaxPool3d and AvgPool3d support [#5098](https://github.com/apache/incubator-tvm/pull/5098) | |
| - Fix build error of iOS RPC [#5621](https://github.com/apache/incubator-tvm/pull/5621) | |
| - Fix three typos [#5620](https://github.com/apache/incubator-tvm/pull/5620) | |
| - isn't a CallNode sometimes [#5623](https://github.com/apache/incubator-tvm/pull/5623) | |
| - Introduce config to PassContext. [#5631](https://github.com/apache/incubator-tvm/pull/5631) | |
| - CMAKE fix [#5630](https://github.com/apache/incubator-tvm/pull/5630) | |
| - Fix typo in test script [#5635](https://github.com/apache/incubator-tvm/pull/5635) | |
| - Label Pattern Partitions [#5627](https://github.com/apache/incubator-tvm/pull/5627) | |
| - Extend AttrPattern to support CallNode and FunctionNode attributes [#5637](https://github.com/apache/incubator-tvm/pull/5637) | |
| - TFLite QNN Tutorial [#5595](https://github.com/apache/incubator-tvm/pull/5595) | |
| - Increase bss section size. [#5660](https://github.com/apache/incubator-tvm/pull/5660) | |
| - Upgrade XGBoost to latest [#5658](https://github.com/apache/incubator-tvm/pull/5658) | |
| - Add buffer name when creating tensor bindings [#5670](https://github.com/apache/incubator-tvm/pull/5670) | |
| - µtvm debug improvements [#5648](https://github.com/apache/incubator-tvm/pull/5648) | |
| - enable amd_apu device on vulkan target [#5659](https://github.com/apache/incubator-tvm/pull/5659) | |
| - Support TupleWrapper as direct ancestor of control flow ops [#5639](https://github.com/apache/incubator-tvm/pull/5639) | |
| - add tvm.micro pydoc to sphinx [#5661](https://github.com/apache/incubator-tvm/pull/5661) | |
| - Add a regression testcase for #5674 [#5677](https://github.com/apache/incubator-tvm/pull/5677) | |
| - Fix C++ RPC build problem on Linux [#5671](https://github.com/apache/incubator-tvm/pull/5671) | |
| - Misc doc fix [#5672](https://github.com/apache/incubator-tvm/pull/5672) | |
| - Add a check Callback to the Pattern Paritioner [#5646](https://github.com/apache/incubator-tvm/pull/5646) | |
| - Call previous excepthook in tvm_excepthook. [#5675](https://github.com/apache/incubator-tvm/pull/5675) | |
| - Fix the shift column for scale_shift_nchw and scale_shift_nhwc in C topi [#5679](https://github.com/apache/incubator-tvm/pull/5679) | |
| - Support more dtypes for TVMDSOOp [#5694](https://github.com/apache/incubator-tvm/pull/5694) | |
| - fix typo: anchor windoes should be anchor windows [#5706](https://github.com/apache/incubator-tvm/pull/5706) | |
| - Remove deprecated opengl files [#5711](https://github.com/apache/incubator-tvm/pull/5711) | |
| - Remove opengl runtime and cmake [#5712](https://github.com/apache/incubator-tvm/pull/5712) | |
| - In memory_plan, check if value is not None, instead of just checking value as boolean. [#5700](https://github.com/apache/incubator-tvm/pull/5700) | |
| - Rename tvm_dso_op to libtvm_dso_op [#5714](https://github.com/apache/incubator-tvm/pull/5714) | |
| - Unify StrMapNode and MapNode [#5687](https://github.com/apache/incubator-tvm/pull/5687) | |
| - Introduce runtime::String::CanConvertFrom [#5718](https://github.com/apache/incubator-tvm/pull/5718) | |
| - Restore the StrMap behavior in JSON/SHash/SEqual [#5719](https://github.com/apache/incubator-tvm/pull/5719) | |
| - Fix generating types like float44 and float88 [#5722](https://github.com/apache/incubator-tvm/pull/5722) | |
| - Avoid downloading when TOPHUB_LOCATION is NONE [#5720](https://github.com/apache/incubator-tvm/pull/5720) | |
| - codegen llvm: move nvptx-specific intrinsic handling into codegen_nvptx [#5726](https://github.com/apache/incubator-tvm/pull/5726) | |
| - ROCm warp shuffles and reductions [#5727](https://github.com/apache/incubator-tvm/pull/5727) | |
| - / delete mismatches in Relay VM [#5735](https://github.com/apache/incubator-tvm/pull/5735) | |
| - Fix flaky test_topi_pooling.py:test_adaptive_pool [#5736](https://github.com/apache/incubator-tvm/pull/5736) | |
| - Fix the values for test_fmod since it fails way too often otherwise [#5723](https://github.com/apache/incubator-tvm/pull/5723) | |
| - fix small bug about dense_grad [#5695](https://github.com/apache/incubator-tvm/pull/5695) | |
| - Fix sequential cpp test [#5745](https://github.com/apache/incubator-tvm/pull/5745) | |
| - Add Scatter to Topi/Relay/ONNX via hybrid script [#5619](https://github.com/apache/incubator-tvm/pull/5619) | |
| - Clean WASM environment before build [#5759](https://github.com/apache/incubator-tvm/pull/5759) | |
| - Second stage of Rust Refactor [#5527](https://github.com/apache/incubator-tvm/pull/5527) | |
| - Fix gelu in PyTorch frontend, tighten numerical checks [#5763](https://github.com/apache/incubator-tvm/pull/5763) | |
| - Make batch matrix multiplication on GPU tunable [#5752](https://github.com/apache/incubator-tvm/pull/5752) | |
| - fix #5686: remove a overstrict assert in MakeAllreduce (#5686) [#5785](https://github.com/apache/incubator-tvm/pull/5785) | |
| - CoreML codegen [#5634](https://github.com/apache/incubator-tvm/pull/5634) | |
| - update vulkan build rule [#5777](https://github.com/apache/incubator-tvm/pull/5777) | |
| - Fix some typos in git-clang-format.sh [#5786](https://github.com/apache/incubator-tvm/pull/5786) | |
| - Edit onnx parser to infer values in post order [#5755](https://github.com/apache/incubator-tvm/pull/5755) | |
| - Support symbolic inputs of Fill [#5762](https://github.com/apache/incubator-tvm/pull/5762) | |
| - support aten::type_as in the pytorch frontend [#5787](https://github.com/apache/incubator-tvm/pull/5787) | |
| - Temporary disable fp16 type_as test for PyTorch Frontend [#5799](https://github.com/apache/incubator-tvm/pull/5799) | |
| - Add config switch for nn.dense layer type. [#5801](https://github.com/apache/incubator-tvm/pull/5801) | |
| - Pin hand landmark network to version 0.7.4. [#5813](https://github.com/apache/incubator-tvm/pull/5813) | |
| - Siju Samuel -> Committer [#5817](https://github.com/apache/incubator-tvm/pull/5817) | |
| - Improve Pattern Language Docs [#5676](https://github.com/apache/incubator-tvm/pull/5676) | |
| - Error msg update [#5818](https://github.com/apache/incubator-tvm/pull/5818) | |
| ## June 2020 | |
| ### Arith | |
| - ExtendedEuclidean merge impl to int_operator [#5625](https://github.com/apache/incubator-tvm/pull/5625) | |
| - Rewrite simplify fix for Vectorized Cooperative Fetching [#5924](https://github.com/apache/incubator-tvm/pull/5924) | |
| ### Fixes | |
| - fix typo: anchor windoes should be anchor windows [#5706](https://github.com/apache/incubator-tvm/pull/5706) | |
| - ReplicationPad support added [#5708](https://github.com/apache/incubator-tvm/pull/5708) | |
| - Simplify Pattern API Implementations [#5703](https://github.com/apache/incubator-tvm/pull/5703) | |
| - Remove deprecated opengl files [#5711](https://github.com/apache/incubator-tvm/pull/5711) | |
| - Remove opengl runtime and cmake [#5712](https://github.com/apache/incubator-tvm/pull/5712) | |
| - Rename tvm_dso_op to libtvm_dso_op [#5714](https://github.com/apache/incubator-tvm/pull/5714) | |
| - Unify StrMapNode and MapNode [#5687](https://github.com/apache/incubator-tvm/pull/5687) | |
| - Introduce runtime::String::CanConvertFrom [#5718](https://github.com/apache/incubator-tvm/pull/5718) | |
| - Restore the StrMap behavior in JSON/SHash/SEqual [#5719](https://github.com/apache/incubator-tvm/pull/5719) | |
| - Fix generating types like float44 and float88 [#5722](https://github.com/apache/incubator-tvm/pull/5722) | |
| - Avoid downloading when TOPHUB_LOCATION is NONE [#5720](https://github.com/apache/incubator-tvm/pull/5720) | |
| - codegen llvm: move nvptx-specific intrinsic handling into codegen_nvptx [#5726](https://github.com/apache/incubator-tvm/pull/5726) | |
| - ROCm warp shuffles and reductions [#5727](https://github.com/apache/incubator-tvm/pull/5727) | |
| - / delete mismatches in Relay VM [#5735](https://github.com/apache/incubator-tvm/pull/5735) | |
| - Fix flaky test_topi_pooling.py:test_adaptive_pool [#5736](https://github.com/apache/incubator-tvm/pull/5736) | |
| - Fix the values for test_fmod since it fails way too often otherwise [#5723](https://github.com/apache/incubator-tvm/pull/5723) | |
| - fix small bug about dense_grad [#5695](https://github.com/apache/incubator-tvm/pull/5695) | |
| - Clarify downstream consistency of TVMArgTypeCode [#5742](https://github.com/apache/incubator-tvm/pull/5742) | |
| - Add Scatter to Topi/Relay/ONNX via hybrid script [#5619](https://github.com/apache/incubator-tvm/pull/5619) | |
| - Clean WASM environment before build [#5759](https://github.com/apache/incubator-tvm/pull/5759) | |
| - Second stage of Rust Refactor [#5527](https://github.com/apache/incubator-tvm/pull/5527) | |
| - Fix gelu in PyTorch frontend, tighten numerical checks [#5763](https://github.com/apache/incubator-tvm/pull/5763) | |
| - Add ShapePattern and DataTypePattern [#5760](https://github.com/apache/incubator-tvm/pull/5760) | |
| - Make batch matrix multiplication on GPU tunable [#5752](https://github.com/apache/incubator-tvm/pull/5752) | |
| - fix #5686: remove a overstrict assert in MakeAllreduce (#5686) [#5785](https://github.com/apache/incubator-tvm/pull/5785) | |
| - CoreML codegen [#5634](https://github.com/apache/incubator-tvm/pull/5634) | |
| - update vulkan build rule [#5777](https://github.com/apache/incubator-tvm/pull/5777) | |
| - aten::norm support added [#5776](https://github.com/apache/incubator-tvm/pull/5776) | |
| - @wpan11nv -> Reviewer [#5790](https://github.com/apache/incubator-tvm/pull/5790) | |
| - Edit onnx parser to infer values in post order [#5755](https://github.com/apache/incubator-tvm/pull/5755) | |
| - Support symbolic inputs of Fill [#5762](https://github.com/apache/incubator-tvm/pull/5762) | |
| - support aten::type_as in the pytorch frontend [#5787](https://github.com/apache/incubator-tvm/pull/5787) | |
| - Temporary disable fp16 type_as test for PyTorch Frontend [#5799](https://github.com/apache/incubator-tvm/pull/5799) | |
| - Add config switch for nn.dense layer type. [#5801](https://github.com/apache/incubator-tvm/pull/5801) | |
| - Move cpu-only frontend tests to a CPU stage [#5807](https://github.com/apache/incubator-tvm/pull/5807) | |
| - Pin hand landmark network to version 0.7.4. [#5813](https://github.com/apache/incubator-tvm/pull/5813) | |
| - Limit number of threads in all jobs [#5815](https://github.com/apache/incubator-tvm/pull/5815) | |
| - Siju Samuel -> Committer [#5817](https://github.com/apache/incubator-tvm/pull/5817) | |
| - Error msg update [#5818](https://github.com/apache/incubator-tvm/pull/5818) | |
| - fix relay.build to not change the module argument in place [#5822](https://github.com/apache/incubator-tvm/pull/5822) | |
| - Fix InferType when module contains Prelude [#5797](https://github.com/apache/incubator-tvm/pull/5797) | |
| - Fix v0.6 CI [#5832](https://github.com/apache/incubator-tvm/pull/5832) | |
| - Add a combine batch_matmul pass [#5791](https://github.com/apache/incubator-tvm/pull/5791) | |
| - RepeatVector, Conv3DTranspose op support added [#5833](https://github.com/apache/incubator-tvm/pull/5833) | |
| - Fix converting serialized quantized models [#5839](https://github.com/apache/incubator-tvm/pull/5839) | |
| - ffi (Object): make class dict visible in instances [#5843](https://github.com/apache/incubator-tvm/pull/5843) | |
| - `tvm` crate stage 3 of Rust refactor [#5769](https://github.com/apache/incubator-tvm/pull/5769) | |
| - Additional canonicalization added for AddNode [#5846](https://github.com/apache/incubator-tvm/pull/5846) | |
| - Suppress the warning messages when compile engine selects impls [#5821](https://github.com/apache/incubator-tvm/pull/5821) | |
| - fix #5849 [#5851](https://github.com/apache/incubator-tvm/pull/5851) | |
| - Introduce POD-C Compliant tvm::Map [#5740](https://github.com/apache/incubator-tvm/pull/5740) | |
| - Add bfloat16 [#5601](https://github.com/apache/incubator-tvm/pull/5601) | |
| - Add Python Classes for all Attrs [#5853](https://github.com/apache/incubator-tvm/pull/5853) | |
| - Fix map assign issue in CI test [#5854](https://github.com/apache/incubator-tvm/pull/5854) | |
| - Introduce Target Id Registry [#5838](https://github.com/apache/incubator-tvm/pull/5838) | |
| - Update has_dtype/has_shape to pattern lang doc [#5847](https://github.com/apache/incubator-tvm/pull/5847) | |
| - Add nn.batch_flatten as quantizable. [#5805](https://github.com/apache/incubator-tvm/pull/5805) | |
| - Fail early before running invalid dynamic graphs [#5856](https://github.com/apache/incubator-tvm/pull/5856) | |
| - Improve type handling in PyTorch frontend [#5834](https://github.com/apache/incubator-tvm/pull/5834) | |
| - Matthew Brookhart -> Reviewer [#5886](https://github.com/apache/incubator-tvm/pull/5886) | |
| - keep parameter names from PyTorch [#5887](https://github.com/apache/incubator-tvm/pull/5887) | |
| - Improve quantized convolution performance for armv8 architectures [#5754](https://github.com/apache/incubator-tvm/pull/5754) | |
| - HotFix the python intrin rule [#5895](https://github.com/apache/incubator-tvm/pull/5895) | |
| - Rust Refactor Stage 4: Rewrite Rust graph runtime to use new APIs [#5830](https://github.com/apache/incubator-tvm/pull/5830) | |
| - add a few gradients [#5899](https://github.com/apache/incubator-tvm/pull/5899) | |
| - Add Binary Intrinsic ops to TIR Ops in C++ [#5900](https://github.com/apache/incubator-tvm/pull/5900) | |
| - Allow implicit conversion in TVM FFI to tvm::Bool [#5907](https://github.com/apache/incubator-tvm/pull/5907) | |
| - PyTorch frontend: fix handling of duplicate use of a model weight [#5897](https://github.com/apache/incubator-tvm/pull/5897) | |
| - Don't multiply by constant 1 uselessly in dense [#5911](https://github.com/apache/incubator-tvm/pull/5911) | |
| - Support any index matching for TupleGetItem [#5909](https://github.com/apache/incubator-tvm/pull/5909) | |
| - Add MicroTVM tutorial using the STM32F746 discovery board [#5655](https://github.com/apache/incubator-tvm/pull/5655) | |
| - Fix serialization of inf float value [#5912](https://github.com/apache/incubator-tvm/pull/5912) | |
| - Fix CPU Thread Binding for Multiple Sockets [#5918](https://github.com/apache/incubator-tvm/pull/5918) | |
| - CUDA device API & VerifyGPUCode pass update [#5898](https://github.com/apache/incubator-tvm/pull/5898) | |
| - Update install.rst [#5858](https://github.com/apache/incubator-tvm/pull/5858) | |
| - Two small fixes to AMDCPU codegen for LLVM 10+ and ROCm 3.5+ [#5920](https://github.com/apache/incubator-tvm/pull/5920) | |
| - Add LegalizeInvalidAttach to legalize the compute_at location after split or fuse [#5917](https://github.com/apache/incubator-tvm/pull/5917) | |
| - Update code_review.rst [#5923](https://github.com/apache/incubator-tvm/pull/5923) | |
| - Don't rewrite expressions used outside of the pattern [#5930](https://github.com/apache/incubator-tvm/pull/5930) | |
| - Add TupleGetItem to CSE [#5931](https://github.com/apache/incubator-tvm/pull/5931) | |
| - Various update for CoreML codegen [#5934](https://github.com/apache/incubator-tvm/pull/5934) | |
| - Update date in the NOTICE [#5943](https://github.com/apache/incubator-tvm/pull/5943) | |
| - Update date in the NOTICE [#5942](https://github.com/apache/incubator-tvm/pull/5942) | |
| - minor fix for release doc [#5948](https://github.com/apache/incubator-tvm/pull/5948) | |
| - raise right error in tensorflow split op [#5951](https://github.com/apache/incubator-tvm/pull/5951) | |
| - add rm xla attributes in tf docs [#5950](https://github.com/apache/incubator-tvm/pull/5950) | |
| - Fix some typo errors in license header [#5957](https://github.com/apache/incubator-tvm/pull/5957) | |
| - Fix OpenCL get_valid_counts errors due to intrinsic atomic_add [#5857](https://github.com/apache/incubator-tvm/pull/5857) | |
| - Fix some typo errors in license header [#5956](https://github.com/apache/incubator-tvm/pull/5956) | |
| - Amendments for gradients [#5941](https://github.com/apache/incubator-tvm/pull/5941) | |
| - Fix the meaning of conv{1,2}d_transpose output_padding parameter. [#5758](https://github.com/apache/incubator-tvm/pull/5758) | |
| - Make first order gradient graphs more efficient [#5959](https://github.com/apache/incubator-tvm/pull/5959) | |
| - Raise an exception when extern function does not return Stmt [#5964](https://github.com/apache/incubator-tvm/pull/5964) | |
| - Fix small typo in nn.conv2d_gemm_weight_transform [#5925](https://github.com/apache/incubator-tvm/pull/5925) | |
| - Improve docker/bash.sh to handle git worktrees [#5970](https://github.com/apache/incubator-tvm/pull/5970) | |
| - Install DNNL (OneDNN) to CI Environment [#5936](https://github.com/apache/incubator-tvm/pull/5936) | |
| - Add Dynamic reshape to a dynamic namespace and add DynamicToStatic Pass [#5826](https://github.com/apache/incubator-tvm/pull/5826) | |
| - Add meshgrid op in Relay, TOPI, Pytorch frontend [#5961](https://github.com/apache/incubator-tvm/pull/5961) | |
| - Print right number of parentheses for LoadNode [#5965](https://github.com/apache/incubator-tvm/pull/5965) | |
| - fix tvm relay testing tf.py typo error [#5977](https://github.com/apache/incubator-tvm/pull/5977) | |
| - Migrate data structure of TargetNode [#5960](https://github.com/apache/incubator-tvm/pull/5960) | |
| - Remove redundant function CreateBufferVecPtr [#5982](https://github.com/apache/incubator-tvm/pull/5982) | |
| - Fix string argument mismatch in GraphRuntimeCodegen [#5933](https://github.com/apache/incubator-tvm/pull/5933) | |
| - Demo showing how to run a pruned 🤗 model. [#5975](https://github.com/apache/incubator-tvm/pull/5975) | |
| - VectorType::get with two parameters is deprecated in LLVM 11+ [#5984](https://github.com/apache/incubator-tvm/pull/5984) | |
| ### Refactor | |
| - relay.op.Op -> tvm.ir.Op [#5705](https://github.com/apache/incubator-tvm/pull/5705) | |
| - Separate ArgTypeCode from DLDataTypeCode [#5730](https://github.com/apache/incubator-tvm/pull/5730) | |
| - Remove legacy compute_expr.h [#5738](https://github.com/apache/incubator-tvm/pull/5738) | |
| - Call::Halide => ProducerLoad, DSL/TIR decouple. [#5743](https://github.com/apache/incubator-tvm/pull/5743) | |
| - Provide->ProducerStore, Realize->ProducerRealize. [#5750](https://github.com/apache/incubator-tvm/pull/5750) | |
| - Migrate the tvm/tir/expr.h to constructor [#5773](https://github.com/apache/incubator-tvm/pull/5773) | |
| - Migrate tir/stmt.h to use constructor. [#5778](https://github.com/apache/incubator-tvm/pull/5778) | |
| - Migrate all Object construction to constructor. [#5784](https://github.com/apache/incubator-tvm/pull/5784) | |
| - Cleanup unused classes [#5789](https://github.com/apache/incubator-tvm/pull/5789) | |
| - Finish std::string->String updates [#5793](https://github.com/apache/incubator-tvm/pull/5793) | |
| - Add tir prefix to type keys [#5802](https://github.com/apache/incubator-tvm/pull/5802) | |
| - Deprecate FreeStmt [#5890](https://github.com/apache/incubator-tvm/pull/5890) | |
| - Change Call.name to Call.op(RelayExpr) [#5863](https://github.com/apache/incubator-tvm/pull/5863) | |
| - Range/IntSet API style consistency. [#5953](https://github.com/apache/incubator-tvm/pull/5953) | |
| ### Bugfix | |
| - Fix Compilation Error in CRT [#5713](https://github.com/apache/incubator-tvm/pull/5713) | |
| - Fix runtime::String backward compatibility in JSON [#5725](https://github.com/apache/incubator-tvm/pull/5725) | |
| - Allow RPCWrappedFunc to rewrite runtime::String as std::string [#5796](https://github.com/apache/incubator-tvm/pull/5796) | |
| - Fix reshape [#5739](https://github.com/apache/incubator-tvm/pull/5739) | |
| - Fix building with LLVM-10 on macOS [#5859](https://github.com/apache/incubator-tvm/pull/5859) | |
| - Add cuda 11 to contrib.nvcc.find_libdevice_path() [#5902](https://github.com/apache/incubator-tvm/pull/5902) | |
| ### Mxnet | |
| - Softmin, trunc op support added [#5715](https://github.com/apache/incubator-tvm/pull/5715) | |
| - conv3d and conv3d_transpose addedx [#5814](https://github.com/apache/incubator-tvm/pull/5814) | |
| - Add parser for contrib.box_decode [#5967](https://github.com/apache/incubator-tvm/pull/5967) | |
| ### Onnx | |
| - ReduceL1, ReduceL2, ReduceSumSquare, ReduceLogSum ops added [#5721](https://github.com/apache/incubator-tvm/pull/5721) | |
| - MaxRoiPool, Mod & Xor op support added [#5729](https://github.com/apache/incubator-tvm/pull/5729) | |
| - Skip multiply with 1.0f constant for GEMM import [#5800](https://github.com/apache/incubator-tvm/pull/5800) | |
| - Fix an issue with #5755 and add Batch norm unit tests. [#5845](https://github.com/apache/incubator-tvm/pull/5845) | |
| ### Tensorflow | |
| - StatefulPartitionedCall/PartitionedCall Ops support added [#5617](https://github.com/apache/incubator-tvm/pull/5617) | |
| - Don't add cast for batch norm when type isn't changing [#5731](https://github.com/apache/incubator-tvm/pull/5731) | |
| - Conv3d Transpose OP added [#5775](https://github.com/apache/incubator-tvm/pull/5775) | |
| ### Relay | |
| - Clear compile engine after task extraction [#5724](https://github.com/apache/incubator-tvm/pull/5724) | |
| - Sparse to dense operator [#5447](https://github.com/apache/incubator-tvm/pull/5447) | |
| - support dynamic NMS(Non Maximum Suppression), symbolic begin, end, and strides for strided_slice [#4312](https://github.com/apache/incubator-tvm/pull/4312) | |
| - Conv3d_transpose op support added [#5737](https://github.com/apache/incubator-tvm/pull/5737) | |
| - Fix for recursive let [#5757](https://github.com/apache/incubator-tvm/pull/5757) | |
| - Fix Calibration Pass to Support Modules with Multiple Functions [#5768](https://github.com/apache/incubator-tvm/pull/5768) | |
| - Add storage_order ignore in pooling layer. [#5781](https://github.com/apache/incubator-tvm/pull/5781) | |
| - Tweak cublas/cudnn priority level [#5820](https://github.com/apache/incubator-tvm/pull/5820) | |
| - ReverseSequence operator [#5495](https://github.com/apache/incubator-tvm/pull/5495) | |
| - Add operation gather to relay. [#5716](https://github.com/apache/incubator-tvm/pull/5716) | |
| - Skip Unknown Function Symbols [#5888](https://github.com/apache/incubator-tvm/pull/5888) | |
| - Allow every runtime module to handle constants [#5885](https://github.com/apache/incubator-tvm/pull/5885) | |
| - Some performance improvement to VM [#5901](https://github.com/apache/incubator-tvm/pull/5901) | |
| - Add shape_of instruction [#5855](https://github.com/apache/incubator-tvm/pull/5855) | |
| - symbolic max_output_size [#5844](https://github.com/apache/incubator-tvm/pull/5844) | |
| - handle Tuple/TupleGetItem in first order gradient [#5946](https://github.com/apache/incubator-tvm/pull/5946) | |
| - Add resnet-3d & Update network definitions for NHWC layout [#5945](https://github.com/apache/incubator-tvm/pull/5945) | |
| ### Frontend | |
| - Add parser support for shape and range [#5329](https://github.com/apache/incubator-tvm/pull/5329) | |
| - Darknet support batch size for yolo [#5688](https://github.com/apache/incubator-tvm/pull/5688) | |
| - Improve Control Flow and TensorArray [#5699](https://github.com/apache/incubator-tvm/pull/5699) | |
| - Improve TF Parser to keep output nodes for saved_model [#5794](https://github.com/apache/incubator-tvm/pull/5794) | |
| - Add parser support for relu6, leaky_relu, relu_n1_to_1, log_softmax [#4805](https://github.com/apache/incubator-tvm/pull/4805) | |
| - Fix TF Dynamic input shape [#5825](https://github.com/apache/incubator-tvm/pull/5825) | |
| - Support a few contrib ops in mxnet [#5819](https://github.com/apache/incubator-tvm/pull/5819) | |
| - Check all unsupported ops before raising an exception [#5929](https://github.com/apache/incubator-tvm/pull/5929) | |
| # Topi | |
| - Fix reshape usage in ARM schedule [#5732](https://github.com/apache/incubator-tvm/pull/5732) | |
| - block sparse dense on cuda [#5746](https://github.com/apache/incubator-tvm/pull/5746) | |
| - pass-by-value -> pass-by-const-reference [#5783](https://github.com/apache/incubator-tvm/pull/5783) | |
| - fix sparse dense schedule on cuda [#5803](https://github.com/apache/incubator-tvm/pull/5803) | |
| - fix strategy for sparse dense cuda [#5782](https://github.com/apache/incubator-tvm/pull/5782) | |
| - Fix x86 conv2d template when tuning with unpacked layout [#5938](https://github.com/apache/incubator-tvm/pull/5938) | |
| ### Fix | |
| - Fix sequential cpp test [#5745](https://github.com/apache/incubator-tvm/pull/5745) | |
| - Infer types in MergeComposite [#5766](https://github.com/apache/incubator-tvm/pull/5766) | |
| - Fix some typos in git-clang-format.sh [#5786](https://github.com/apache/incubator-tvm/pull/5786) | |
| - Fix recursive let for well formed check [#5780](https://github.com/apache/incubator-tvm/pull/5780) | |
| - Recover global state after test_util.py [#5824](https://github.com/apache/incubator-tvm/pull/5824) | |
| ### Backport-0.6 | |
| - fix a min/max simplify bug [#5749](https://github.com/apache/incubator-tvm/pull/5749) | |
| - fix a min/max simplify bug [#5761](https://github.com/apache/incubator-tvm/pull/5761) | |
| - Fix alpha_equal bug [#5829](https://github.com/apache/incubator-tvm/pull/5829) | |
| - fix RemoveUnusedFunctions pass [#5828](https://github.com/apache/incubator-tvm/pull/5828) | |
| - Add ConstantNode to IsAtomic [#5831](https://github.com/apache/incubator-tvm/pull/5831) | |
| - Fix search path for libtvm_topi.so [#5836](https://github.com/apache/incubator-tvm/pull/5836) | |
| - Fix Python debugger segfaults with TVM built with LLVM [#5837](https://github.com/apache/incubator-tvm/pull/5837) | |
| - Fixed process termination routine in windows [#5849](https://github.com/apache/incubator-tvm/pull/5849) | |
| - Fix annotation for multiply op (#4458) [#5850](https://github.com/apache/incubator-tvm/pull/5850) | |
| - Fix NDArray SaveDLTensor declaration and implementation signature different [#5852](https://github.com/apache/incubator-tvm/pull/5852) | |
| - fix serialization precision loss in float [#5860](https://github.com/apache/incubator-tvm/pull/5860) | |
| - fix _parse_param bug [#5861](https://github.com/apache/incubator-tvm/pull/5861) | |
| - Fix bias_add gradient [#5862](https://github.com/apache/incubator-tvm/pull/5862) | |
| - Make sure to visit the arguments of inlined functions [#5864](https://github.com/apache/incubator-tvm/pull/5864) | |
| - Fix Python syntax error in start_rpc_server_to_tracker.py [#5865](https://github.com/apache/incubator-tvm/pull/5865) | |
| - Fixed crash caused by reversing bitwise operations [#5866](https://github.com/apache/incubator-tvm/pull/5866) | |
| - Fix copy constructor [#5867](https://github.com/apache/incubator-tvm/pull/5867) | |
| - fix small bug about dense_grad [#5868](https://github.com/apache/incubator-tvm/pull/5868) | |
| - Fix compile errors of OpenCL FPGA backend [#5869](https://github.com/apache/incubator-tvm/pull/5869) | |
| - Some Windows and MSVC fixes [#5870](https://github.com/apache/incubator-tvm/pull/5870) | |
| - LRN only supports 4D tensors, remove it from alter_op [#5871](https://github.com/apache/incubator-tvm/pull/5871) | |
| - fix topi.nn.global_pool layout=NHWC [#5872](https://github.com/apache/incubator-tvm/pull/5872) | |
| - Fix hasattr by extracting Python error type from Windows error message [#5873](https://github.com/apache/incubator-tvm/pull/5873) | |
| - Export GraphRuntime in tvm_runtime.dll [#5874](https://github.com/apache/incubator-tvm/pull/5874) | |
| - Fix Base64OutStream portability issue [#5875](https://github.com/apache/incubator-tvm/pull/5875) | |
| - Fix a bug in generating the search space [#5876](https://github.com/apache/incubator-tvm/pull/5876) | |
| - Fix compilation of If-Elses [#5877](https://github.com/apache/incubator-tvm/pull/5877) | |
| - Fix FuseBatchNorm output cast error if need_cast is True [#5878](https://github.com/apache/incubator-tvm/pull/5878) | |
| - fskip of EliminateCommonSubexpr cannot always return false [#5879](https://github.com/apache/incubator-tvm/pull/5879) | |
| - Fix multiple transfer issue in LoadUop module [#5882](https://github.com/apache/incubator-tvm/pull/5882) | |
| - Enable streamlined GEMM execution [#5893](https://github.com/apache/incubator-tvm/pull/5893) | |
| - Fixed a crash issue in TSIM driver [#5894](https://github.com/apache/incubator-tvm/pull/5894) | |
| - Fix lambda lift pass for recursive call [#5903](https://github.com/apache/incubator-tvm/pull/5903) | |
| - Fix conv2d alter op for arm cpu [#5906](https://github.com/apache/incubator-tvm/pull/5906) | |
| - Fix alter op layout when calling a global var [#5904](https://github.com/apache/incubator-tvm/pull/5904) | |
| - Fix dense x86 schedule [#5905](https://github.com/apache/incubator-tvm/pull/5905) | |
| - End-to-end Inference with Chisel VTA [#5896](https://github.com/apache/incubator-tvm/pull/5896) | |
| - keep div_mode during floordiv simplify [#5927](https://github.com/apache/incubator-tvm/pull/5927) | |
| - keep div_mode during floordiv simplify [#5922](https://github.com/apache/incubator-tvm/pull/5922) | |
| - fskip of EliminateCommonSubexpr cannot always return false [#5880](https://github.com/apache/incubator-tvm/pull/5880) | |
| ### Runtime | |
| - Add compile_shared option to linux compile utility fn [#5751](https://github.com/apache/incubator-tvm/pull/5751) | |
| - Overload string operators [#5806](https://github.com/apache/incubator-tvm/pull/5806) | |
| - Introduce MetadataModule to separate code compilation/interpretation and weight initialization [#5770](https://github.com/apache/incubator-tvm/pull/5770) | |
| - Only initialize required module [#5926](https://github.com/apache/incubator-tvm/pull/5926) | |
| ### Tir | |
| - Remove CallNode.call_type in favor of attribute. [#5937](https://github.com/apache/incubator-tvm/pull/5937) | |
| - Remove legacy HoistIfThenElse [#5944](https://github.com/apache/incubator-tvm/pull/5944) | |
| - Improve Let/LetStmt support. [#5949](https://github.com/apache/incubator-tvm/pull/5949) | |
| - Refine side effect analysis. [#5954](https://github.com/apache/incubator-tvm/pull/5954) | |
| ### Tflite | |
| - QNN support for TFLite 2.1.0 quantized models [#5848](https://github.com/apache/incubator-tvm/pull/5848) | |
| ## July 2020 | |
| ### Topi | |
| - Fix x86 conv2d template when tuning with unpacked layout [#5938](https://github.com/apache/incubator-tvm/pull/5938) | |
| - Fix the filter width parameter in depthwise_conv2d [#6081](https://github.com/apache/incubator-tvm/pull/6081) | |
| - Using MKL blas for quantized dense [#6115](https://github.com/apache/incubator-tvm/pull/6115) | |
| - Fix conv2d_transpose output padding [#6236](https://github.com/apache/incubator-tvm/pull/6236) | |
| - topi -> tvm/topi [#6186](https://github.com/apache/incubator-tvm/pull/6186) | |
| ### Target | |
| - Migrate data structure of TargetNode [#5960](https://github.com/apache/incubator-tvm/pull/5960) | |
| - Use TargetNode::attrs for Target serialization [#5993](https://github.com/apache/incubator-tvm/pull/5993) | |
| - each option of target str should only contain one '=' [#5988](https://github.com/apache/incubator-tvm/pull/5988) | |
| - ONNX codegen [#5052](https://github.com/apache/incubator-tvm/pull/5052) | |
| - Rename target_id => target_kind [#6199](https://github.com/apache/incubator-tvm/pull/6199) | |
| - 64-bit RPi4b target [#6211](https://github.com/apache/incubator-tvm/pull/6211) | |
| ### Relay | |
| - Add resnet-3d & Update network definitions for NHWC layout [#5945](https://github.com/apache/incubator-tvm/pull/5945) | |
| - Small bug fix for Conv1D imports. [#5995](https://github.com/apache/incubator-tvm/pull/5995) | |
| - Fix what looks like bizzare copy-paste issue [#6010](https://github.com/apache/incubator-tvm/pull/6010) | |
| - Add Parser 2.0 [#5932](https://github.com/apache/incubator-tvm/pull/5932) | |
| - Dynamic TopK Op [#6008](https://github.com/apache/incubator-tvm/pull/6008) | |
| - Move invoke_tvm_op and shape_func to vm dialect [#5958](https://github.com/apache/incubator-tvm/pull/5958) | |
| - GRU Layer Support [#6020](https://github.com/apache/incubator-tvm/pull/6020) | |
| - Add pass for getting calibration data from a relay module [#5997](https://github.com/apache/incubator-tvm/pull/5997) | |
| - Dynamic broadcast_to, zeros, ones [#6007](https://github.com/apache/incubator-tvm/pull/6007) | |
| - Merge two consecutive reshape ops [#6052](https://github.com/apache/incubator-tvm/pull/6052) | |
| - Add operation scatter_add to relay, based on scatter implementation. [#6030](https://github.com/apache/incubator-tvm/pull/6030) | |
| - Add dynamic reshape grad [#6080](https://github.com/apache/incubator-tvm/pull/6080) | |
| - i64 indices [#5235](https://github.com/apache/incubator-tvm/pull/5235) | |
| - Keep fixed dim when unifying dynamic shape [#5795](https://github.com/apache/incubator-tvm/pull/5795) | |
| - Port eliminate_common_subexpr to non-recursive form [#6134](https://github.com/apache/incubator-tvm/pull/6134) | |
| - Fix interpreter for dyanmic shape input of ndarray_size [#6086](https://github.com/apache/incubator-tvm/pull/6086) | |
| - Allow to config allocator type and refactor vm code structure [#6105](https://github.com/apache/incubator-tvm/pull/6105) | |
| - Support NMSv4 [#6085](https://github.com/apache/incubator-tvm/pull/6085) | |
| - Handle ndarray_size in FoldConstant [#6156](https://github.com/apache/incubator-tvm/pull/6156) | |
| - when converting constant nodes with types of int64 or float64 [#6159](https://github.com/apache/incubator-tvm/pull/6159) | |
| - Add ReshapeTensor instruction in the VM to replace the reshape op [#6089](https://github.com/apache/incubator-tvm/pull/6089) | |
| - Fix bug in transpose_shape_func [#6180](https://github.com/apache/incubator-tvm/pull/6180) | |
| - Basic block normal form [#6152](https://github.com/apache/incubator-tvm/pull/6152) | |
| - pytorch frontend support conv1d [#6203](https://github.com/apache/incubator-tvm/pull/6203) | |
| - OneHot operation [#6209](https://github.com/apache/incubator-tvm/pull/6209) | |
| - Support combine multiple dense op just into dense [#6062](https://github.com/apache/incubator-tvm/pull/6062) | |
| - Add Dynamic Resize Op [#6198](https://github.com/apache/incubator-tvm/pull/6198) | |
| - Add unbiased variance op and corresponding support in pytorch frontend [#6232](https://github.com/apache/incubator-tvm/pull/6232) | |
| - Refine tensorflow frontend 1.x & 2.x compatibility [#6240](https://github.com/apache/incubator-tvm/pull/6240) | |
| ### Runtime | |
| - if a param not in input, we should still consume it's data [#5990](https://github.com/apache/incubator-tvm/pull/5990) | |
| - Support module based interface runtime [#5753](https://github.com/apache/incubator-tvm/pull/5753) | |
| - init TVMPackedFunc's name [#6044](https://github.com/apache/incubator-tvm/pull/6044) | |
| - Enable auto conversion String->DLDataType [#6214](https://github.com/apache/incubator-tvm/pull/6214) | |
| - fix typo [#6230](https://github.com/apache/incubator-tvm/pull/6230) | |
| ### Ci | |
| - Update ci-cpu to the latest [#6031](https://github.com/apache/incubator-tvm/pull/6031) | |
| - Move CI over to new Rust crates and try to fix flaky test. [#6011](https://github.com/apache/incubator-tvm/pull/6011) | |
| - Add ACL docker installation [#5916](https://github.com/apache/incubator-tvm/pull/5916) | |
| - Temporary disable nmsv4 test [#6151](https://github.com/apache/incubator-tvm/pull/6151) | |
| - Update ci-cpu to the latest [#6164](https://github.com/apache/incubator-tvm/pull/6164) | |
| - add caffe environment [#6023](https://github.com/apache/incubator-tvm/pull/6023) | |
| - Remove topi from the CI cache [#6188](https://github.com/apache/incubator-tvm/pull/6188) | |
| - Enable CI for Ethos-N [#6171](https://github.com/apache/incubator-tvm/pull/6171) | |
| ### Byoc | |
| - JSON Runtime with DNNL End-to-End Flow [#5919](https://github.com/apache/incubator-tvm/pull/5919) | |
| - Handle one symbol for each runtime [#5989](https://github.com/apache/incubator-tvm/pull/5989) | |
| - Run accelerator specific optimizations [#6068](https://github.com/apache/incubator-tvm/pull/6068) | |
| - Arm Compute Library integration [#5915](https://github.com/apache/incubator-tvm/pull/5915) | |
| - Support asymmetric per-layer quantized operators [#6109](https://github.com/apache/incubator-tvm/pull/6109) | |
| - Retire the example json runtime [#6177](https://github.com/apache/incubator-tvm/pull/6177) | |
| - json_node.h should include data_type.h [#6224](https://github.com/apache/incubator-tvm/pull/6224) | |
| - Improve installation tutorial [#6170](https://github.com/apache/incubator-tvm/pull/6170) | |
| ### Ansor | |
| - Phase 0: Ansor minimum system for auto schedule generating [#5962](https://github.com/apache/incubator-tvm/pull/5962) | |
| - Phase 1: Access Analyzer [#6103](https://github.com/apache/incubator-tvm/pull/6103) | |
| - Phase 1: Add follow_split and follow_fused_split steps [#6142](https://github.com/apache/incubator-tvm/pull/6142) | |
| - Phase 1: Add pragma/storage_align/rfactor steps [#6141](https://github.com/apache/incubator-tvm/pull/6141) | |
| - Phase 1: Add RPC Runner [#6077](https://github.com/apache/incubator-tvm/pull/6077) | |
| - Phase 1: Add annotation/compute_at/compute_root/compute_inline steps [#6073](https://github.com/apache/incubator-tvm/pull/6073) | |
| - Phase 1: Add cache_read/cache_write steps [#6107](https://github.com/apache/incubator-tvm/pull/6107) | |
| - Phase 1: Rename namspace form auto_schedule to auto_scheduler [#6059](https://github.com/apache/incubator-tvm/pull/6059) | |
| - Phase 1: The base class for cost models [#6187](https://github.com/apache/incubator-tvm/pull/6187) | |
| ### Fix | |
| - Add missing expr visitor for any [#6082](https://github.com/apache/incubator-tvm/pull/6082) | |
| - Remove the tvm web from version update [#6122](https://github.com/apache/incubator-tvm/pull/6122) | |
| - Clear relay cache after every build & Clear warning message cache after autotvm task extraction [#6131](https://github.com/apache/incubator-tvm/pull/6131) | |
| - avoid unexpected throw in AttrInitEntry [#6128](https://github.com/apache/incubator-tvm/pull/6128) | |
| - Verify that tensor reshape is valid. [#6215](https://github.com/apache/incubator-tvm/pull/6215) | |
| ### Docs | |
| - improve the doc of release [#6091](https://github.com/apache/incubator-tvm/pull/6091) | |
| - Cleanup docs build instructions. [#6094](https://github.com/apache/incubator-tvm/pull/6094) | |
| - Organize Design and Architectures [#6097](https://github.com/apache/incubator-tvm/pull/6097) | |
| - Reorganize the docs. [#6146](https://github.com/apache/incubator-tvm/pull/6146) | |
| - Clarify Docs Categorization [#6155](https://github.com/apache/incubator-tvm/pull/6155) | |
| - Improve the docs build instructions [#6173](https://github.com/apache/incubator-tvm/pull/6173) | |
| - Added casting to hybrid script doc and fixed pass infra doc [#6174](https://github.com/apache/incubator-tvm/pull/6174) | |
| - Update pass infra tutorial [#6193](https://github.com/apache/incubator-tvm/pull/6193) | |
| ### Tir | |
| - Improved massive build times caused by tir.floormod and tir.floordiv. Fixed Topi testcase. [#5666](https://github.com/apache/incubator-tvm/pull/5666) | |
| - Buffer logger assert removed [#6147](https://github.com/apache/incubator-tvm/pull/6147) | |
| - Enhance VerifyGPUCode [#6194](https://github.com/apache/incubator-tvm/pull/6194) | |
| - HoistIfThenElse added [#6066](https://github.com/apache/incubator-tvm/pull/6066) | |
| - Hybrid Script Support for TIR [#6227](https://github.com/apache/incubator-tvm/pull/6227) | |
| ### Fixes | |
| - Improve docker/bash.sh to handle git worktrees [#5970](https://github.com/apache/incubator-tvm/pull/5970) | |
| - Add parser for contrib.box_decode [#5967](https://github.com/apache/incubator-tvm/pull/5967) | |
| - Add Dynamic reshape to a dynamic namespace and add DynamicToStatic Pass [#5826](https://github.com/apache/incubator-tvm/pull/5826) | |
| - Add meshgrid op in Relay, TOPI, Pytorch frontend [#5961](https://github.com/apache/incubator-tvm/pull/5961) | |
| - fix tvm relay testing tf.py typo error [#5977](https://github.com/apache/incubator-tvm/pull/5977) | |
| - Remove redundant function CreateBufferVecPtr [#5982](https://github.com/apache/incubator-tvm/pull/5982) | |
| - VectorType::get with two parameters is deprecated in LLVM 11+ [#5984](https://github.com/apache/incubator-tvm/pull/5984) | |
| - QNN support for TFLite 2.1.0 quantized models [#5848](https://github.com/apache/incubator-tvm/pull/5848) | |
| - Inequalities solver [#5618](https://github.com/apache/incubator-tvm/pull/5618) | |
| - Use LocalRunner by default in the tutorial tune_relay_cuda.py [#6001](https://github.com/apache/incubator-tvm/pull/6001) | |
| - Undefined names: import os for line 324 & import re for line 308 [#6003](https://github.com/apache/incubator-tvm/pull/6003) | |
| - GitHub Actions upgrade to actions/setup-python@v2 [#6002](https://github.com/apache/incubator-tvm/pull/6002) | |
| - Dynamic Tile Op [#5983](https://github.com/apache/incubator-tvm/pull/5983) | |
| - Only pass pythonpath for ci images [#6005](https://github.com/apache/incubator-tvm/pull/6005) | |
| - Auto-convert shuffle with single index to "extract element" [#6006](https://github.com/apache/incubator-tvm/pull/6006) | |
| - Cache object refs in loop partitioner instead of object pointers [#6004](https://github.com/apache/incubator-tvm/pull/6004) | |
| - Fix test_arith_solve_linear_inequality.py::test_multi_equal [#6014](https://github.com/apache/incubator-tvm/pull/6014) | |
| - MXNet frontend support for AMP cast op [#5976](https://github.com/apache/incubator-tvm/pull/5976) | |
| - Remove duplicate line [#6017](https://github.com/apache/incubator-tvm/pull/6017) | |
| - Gather op support added [#6013](https://github.com/apache/incubator-tvm/pull/6013) | |
| - Demo showing how to run a pruned 🤗 model. [#5975](https://github.com/apache/incubator-tvm/pull/5975) | |
| - Move compiler related registry items to vta/build_module.py [#6012](https://github.com/apache/incubator-tvm/pull/6012) | |
| - Pin keras version [#6032](https://github.com/apache/incubator-tvm/pull/6032) | |
| - Fix in arm_cpu/conv2d_alter_op for NHWC quantized [#6027](https://github.com/apache/incubator-tvm/pull/6027) | |
| - Add creation of Hexagon device in RPC client [#6035](https://github.com/apache/incubator-tvm/pull/6035) | |
| - Terminate basic block after "ret" instruction [#6036](https://github.com/apache/incubator-tvm/pull/6036) | |
| - µTVM CRT modifications for on-device RPC server [#5921](https://github.com/apache/incubator-tvm/pull/5921) | |
| - Create TBAA information based on the unrelying buffer type [#6046](https://github.com/apache/incubator-tvm/pull/6046) | |
| - Add support for tflite arg_min and arg_max [#5992](https://github.com/apache/incubator-tvm/pull/5992) | |
| - Fix fully_connected converter when batch size is not 1 [#6038](https://github.com/apache/incubator-tvm/pull/6038) | |
| - Fix a primitive check error [#5991](https://github.com/apache/incubator-tvm/pull/5991) | |
| - Refactor to expose MakeOp functions to C++ [#6047](https://github.com/apache/incubator-tvm/pull/6047) | |
| - Fix conv2_gemm after target structure update [#6037](https://github.com/apache/incubator-tvm/pull/6037) | |
| - Remove use of designated initializers from hexagon_module.cc [#6055](https://github.com/apache/incubator-tvm/pull/6055) | |
| - Build crttest and cpptest separately. [#6057](https://github.com/apache/incubator-tvm/pull/6057) | |
| - Fix pytorch frontend prim::Constant issue [#6051](https://github.com/apache/incubator-tvm/pull/6051) | |
| - update frontend tutorials to new model based runtime interface [#6063](https://github.com/apache/incubator-tvm/pull/6063) | |
| - Enable x86 cpu cache flush [#5914](https://github.com/apache/incubator-tvm/pull/5914) | |
| - Remove unnecessary std::cout [#6072](https://github.com/apache/incubator-tvm/pull/6072) | |
| - Fix error message in Buffer::vstore, NFC [#6056](https://github.com/apache/incubator-tvm/pull/6056) | |
| - Fix FSIM Compile Error. [#6070](https://github.com/apache/incubator-tvm/pull/6070) | |
| - Improve vector simplification for float operands [#6043](https://github.com/apache/incubator-tvm/pull/6043) | |
| - Refine LSTMBlockCell to support dynamic rnn [#5963](https://github.com/apache/incubator-tvm/pull/5963) | |
| - Fix LocalBuilder on macOS with python 3.8. [#6083](https://github.com/apache/incubator-tvm/pull/6083) | |
| - Add missing test for fast erf [#6058](https://github.com/apache/incubator-tvm/pull/6058) | |
| - Fixed point multiplication improvements for AArch64 [#5980](https://github.com/apache/incubator-tvm/pull/5980) | |
| - Fix code generation bugs for C/CUDA & Improve VerifyGPUCode pass [#6041](https://github.com/apache/incubator-tvm/pull/6041) | |
| - MXNet pre-quantized BERT [#6039](https://github.com/apache/incubator-tvm/pull/6039) | |
| - Scalar support for `te.extern` [#6079](https://github.com/apache/incubator-tvm/pull/6079) | |
| - Delete declaration of unused op_node [#6102](https://github.com/apache/incubator-tvm/pull/6102) | |
| - Load configs even it has no entity [#6100](https://github.com/apache/incubator-tvm/pull/6100) | |
| - Update SGX example Cargo.toml [#6067](https://github.com/apache/incubator-tvm/pull/6067) | |
| - Add default value for option USE_DNNL_CODEGEN in the cmake [#6099](https://github.com/apache/incubator-tvm/pull/6099) | |
| - Update installation doc with minor improvements [#6104](https://github.com/apache/incubator-tvm/pull/6104) | |
| - lint: add opencl .cl file type [#6092](https://github.com/apache/incubator-tvm/pull/6092) | |
| - Clean up conversions between TVM and Rust functions [#6114](https://github.com/apache/incubator-tvm/pull/6114) | |
| - Improve reduction schedule on arm CPUs [#6110](https://github.com/apache/incubator-tvm/pull/6110) | |
| - Register Shape Func for Some Operators to Handle Dynamic Shapes [#5955](https://github.com/apache/incubator-tvm/pull/5955) | |
| - Fix variable name conflict with OpenCL keyword [#6048](https://github.com/apache/incubator-tvm/pull/6048) | |
| - Some rust cleanups [#6116](https://github.com/apache/incubator-tvm/pull/6116) | |
| - fix typos in comments and relay tutorial [#5999](https://github.com/apache/incubator-tvm/pull/5999) | |
| - Option to specify alternate directory to output build to [#6016](https://github.com/apache/incubator-tvm/pull/6016) | |
| - Add 'get_num_inputs' to GraphRuntime [#6118](https://github.com/apache/incubator-tvm/pull/6118) | |
| - TFLite quantized conv test [#6084](https://github.com/apache/incubator-tvm/pull/6084) | |
| - Fix autotvm on the conv2d_nchw_winograd.mali operator [#6130](https://github.com/apache/incubator-tvm/pull/6130) | |
| - add attr option mfloat-abi for arm32 [#6123](https://github.com/apache/incubator-tvm/pull/6123) | |
| - Fix CUDA Library Tuning [#6132](https://github.com/apache/incubator-tvm/pull/6132) | |
| - Add missing RPC sources after refactor [#6113](https://github.com/apache/incubator-tvm/pull/6113) | |
| - Add TVM application extension with WASM runtime [#5892](https://github.com/apache/incubator-tvm/pull/5892) | |
| - @t-vi -> Reviewer [#6149](https://github.com/apache/incubator-tvm/pull/6149) | |
| - Correct runtime.load_module [#6161](https://github.com/apache/incubator-tvm/pull/6161) | |
| - Improve error messages in graph tuner, graph runtime, and module loader. [#6148](https://github.com/apache/incubator-tvm/pull/6148) | |
| - Typo in mod creation [#6165](https://github.com/apache/incubator-tvm/pull/6165) | |
| - Fix some shape mismatches between TF and Relay [#6166](https://github.com/apache/incubator-tvm/pull/6166) | |
| - Improve doc string [#6176](https://github.com/apache/incubator-tvm/pull/6176) | |
| - Fix incorrect function signature in header [#6172](https://github.com/apache/incubator-tvm/pull/6172) | |
| - Temporary disable conv2d grad strided flaky test [#6183](https://github.com/apache/incubator-tvm/pull/6183) | |
| - Remove libtopi from the build [#6189](https://github.com/apache/incubator-tvm/pull/6189) | |
| - Create Interpreter for each constant subgraph [#6195](https://github.com/apache/incubator-tvm/pull/6195) | |
| - Fix alignment of note [#6181](https://github.com/apache/incubator-tvm/pull/6181) | |
| - Implemented PADV2 Operator for TFLite and added support for constant values in PAD. [#6167](https://github.com/apache/incubator-tvm/pull/6167) | |
| - Unary ops support added in frontend [#6196](https://github.com/apache/incubator-tvm/pull/6196) | |
| - fix #6205 [#6207](https://github.com/apache/incubator-tvm/pull/6207) | |
| - Change the meaning of conv3d_transpose output_padding to match conv{1,2}d_transpose [#6065](https://github.com/apache/incubator-tvm/pull/6065) | |
| - Fix compile warnings. [#6204](https://github.com/apache/incubator-tvm/pull/6204) | |
| - Fix -mfloat-abi=soft compilation for ARM with OpenCL target [#6150](https://github.com/apache/incubator-tvm/pull/6150) | |
| - match pytorch 1.6 googlenet pretrained model (#6201) [#6212](https://github.com/apache/incubator-tvm/pull/6212) | |
| - Add --runtime=c, remove micro_dev target, enable LLVM backend [#6145](https://github.com/apache/incubator-tvm/pull/6145) | |
| - Mod operator, bug fix [#6160](https://github.com/apache/incubator-tvm/pull/6160) | |
| - RESHAPE with dynamic shape arg in TFLite frontend [#6208](https://github.com/apache/incubator-tvm/pull/6208) | |
| - fix compilation error with cuda 11 [#6213](https://github.com/apache/incubator-tvm/pull/6213) | |
| - fix port_end wrong default value 9199 to 9099 for keeping same with source code [#6220](https://github.com/apache/incubator-tvm/pull/6220) | |
| - Std op without specified dimensions support [#6226](https://github.com/apache/incubator-tvm/pull/6226) | |
| - fix crt building and running error [#6231](https://github.com/apache/incubator-tvm/pull/6231) | |
| - jcf94 -> Reviewer [#6241](https://github.com/apache/incubator-tvm/pull/6241) | |
| - Implemented ONE_HOT Operator for TFLite. [#6223](https://github.com/apache/incubator-tvm/pull/6223) | |
| ##Aug 2020 | |
| ### Runtime | |
| - Support random fill [#5913](https://github.com/apache/incubator-tvm/pull/5913) | |
| - Use new to avoid exit-time de-allocation order [#6292](https://github.com/apache/incubator-tvm/pull/6292) | |
| - Add parallel_for support to run a loop in parallel [#6275](https://github.com/apache/incubator-tvm/pull/6275) | |
| ### Tir | |
| - Enhance VerifyGPUCode [#6194](https://github.com/apache/incubator-tvm/pull/6194) | |
| - HoistIfThenElse added [#6066](https://github.com/apache/incubator-tvm/pull/6066) | |
| - Hybrid Script Support for TIR [#6227](https://github.com/apache/incubator-tvm/pull/6227) | |
| - Enforce buffer pointer var type to be consistent with dtype. [#6317](https://github.com/apache/incubator-tvm/pull/6317) | |
| - Block scope hoisting added [#6238](https://github.com/apache/incubator-tvm/pull/6238) | |
| ### Target | |
| - Rename target_id => target_kind [#6199](https://github.com/apache/incubator-tvm/pull/6199) | |
| - 64-bit RPi4b target [#6211](https://github.com/apache/incubator-tvm/pull/6211) | |
| - Creating Target from JSON-like Configuration [#6218](https://github.com/apache/incubator-tvm/pull/6218) | |
| - Add python binding to new JSON target construction [#6315](https://github.com/apache/incubator-tvm/pull/6315) | |
| - Use target class in all codegens [#6347](https://github.com/apache/incubator-tvm/pull/6347) | |
| ### Codegen | |
| - Initial support for Hexagon codegen [#6261](https://github.com/apache/incubator-tvm/pull/6261) | |
| - Add --runtime=c, remove micro_dev target, enable LLVM backend [#6145](https://github.com/apache/incubator-tvm/pull/6145) | |
| - Add tvm::support::hexdump() debug utility [#6154](https://github.com/apache/incubator-tvm/pull/6154) | |
| ### Relay | |
| - Basic block normal form [#6152](https://github.com/apache/incubator-tvm/pull/6152) | |
| - OneHot operation [#6209](https://github.com/apache/incubator-tvm/pull/6209) | |
| - Support combine multiple dense op just into dense [#6062](https://github.com/apache/incubator-tvm/pull/6062) | |
| - Add Dynamic Resize Op [#6198](https://github.com/apache/incubator-tvm/pull/6198) | |
| - Dynamic full operator [#6260](https://github.com/apache/incubator-tvm/pull/6260) | |
| - Fix node indices attribute error for tensorflow 2.3 [#6288](https://github.com/apache/incubator-tvm/pull/6288) | |
| - Make the max number of fused ops configurable [#6327](https://github.com/apache/incubator-tvm/pull/6327) | |
| - Implementation of the dynamic pad operator [#6284](https://github.com/apache/incubator-tvm/pull/6284) | |
| - change device annotation from post DFS to recursive [#6124](https://github.com/apache/incubator-tvm/pull/6124) | |
| - Dynamic upsampling relay op [#6273](https://github.com/apache/incubator-tvm/pull/6273) | |
| - Make check stricter: disallow inserting function with free vars into module [#6313](https://github.com/apache/incubator-tvm/pull/6313) | |
| - Support for PyTorch Non-Maximum Suppression [#6314](https://github.com/apache/incubator-tvm/pull/6314) | |
| - Make check stricter by using Feature. Fixed multiple bugs [#6326](https://github.com/apache/incubator-tvm/pull/6326) | |
| - Resize support for NCHW-convertible layouts [#6293](https://github.com/apache/incubator-tvm/pull/6293) | |
| - Make AutoDiff thread through global function [#6336](https://github.com/apache/incubator-tvm/pull/6336) | |
| - Create Interpreter for each constant subgraph [#6195](https://github.com/apache/incubator-tvm/pull/6195) | |
| - Parser 2.0 part 2 [#6162](https://github.com/apache/incubator-tvm/pull/6162) | |
| ### Ansor | |
| - Phase 1: The base class for cost models [#6187](https://github.com/apache/incubator-tvm/pull/6187) | |
| - Phase 2: Basic CPU Sketch Search Policy [#6184](https://github.com/apache/incubator-tvm/pull/6184) | |
| - Phase 1: feature extraction for cost models [#6190](https://github.com/apache/incubator-tvm/pull/6190) | |
| - Phase 1: XGBoost Cost Model [#6270](https://github.com/apache/incubator-tvm/pull/6270) | |
| - Phase 2: Basic GPU Sketch Search Policy [#6269](https://github.com/apache/incubator-tvm/pull/6269) | |
| - Phase 2: Evolutionary Search [#6310](https://github.com/apache/incubator-tvm/pull/6310) | |
| - Phase 2: Update heavy operations with parallel_for [#6348](https://github.com/apache/incubator-tvm/pull/6348) | |
| ### BYOC | |
| - json_node.h should include data_type.h [#6224](https://github.com/apache/incubator-tvm/pull/6224) | |
| - Improve installation tutorial [#6170](https://github.com/apache/incubator-tvm/pull/6170) | |
| - Add support for dense (fully connected) layer [#6254](https://github.com/apache/incubator-tvm/pull/6254) | |
| - Introduce the Ethos-N BYOC integration [#6222](https://github.com/apache/incubator-tvm/pull/6222) | |
| - Enable remote device via environment variables [#6279](https://github.com/apache/incubator-tvm/pull/6279) | |
| - Improved pooling support [#6248](https://github.com/apache/incubator-tvm/pull/6248) | |
| - Add support for quantized convolution [#6335](https://github.com/apache/incubator-tvm/pull/6335) | |
| ### PyTorch | |
| - Add Pytorch advanced indexing [#6318](https://github.com/apache/incubator-tvm/pull/6318) | |
| - Support index_select [#6295](https://github.com/apache/incubator-tvm/pull/6295) | |
| - Fix cast to long [#6301](https://github.com/apache/incubator-tvm/pull/6301) | |
| - Fix dtype handling for modules with integer parameters [#6311](https://github.com/apache/incubator-tvm/pull/6311) | |
| - pytorch frontend support conv1d [#6203](https://github.com/apache/incubator-tvm/pull/6203) | |
| - Add cast to double, fix flatten conversion [#6357](https://github.com/apache/incubator-tvm/pull/6357) | |
| - Fix aten::max and aten::min conversion [#6372](https://github.com/apache/incubator-tvm/pull/6372) | |
| - Match pytorch 1.6 googlenet pretrained model (#6201) [#6212](https://github.com/apache/incubator-tvm/pull/6212)- Add unbiased variance op and corresponding support in pytorch frontend [#6232](https://github.com/apache/incubator-tvm/pull/6232) | |
| ### TFLite | |
| - Implemented PADV2 Operator for TFLite and added support for constant values in PAD. [#6167](https://github.com/apache/incubator-tvm/pull/6167) | |
| - Implemented ONE_HOT Operator for TFLite. [#6223](https://github.com/apache/incubator-tvm/pull/6223) | |
| - Implemented EXPAND_DIMS Operator for TFLite. [#6243](https://github.com/apache/incubator-tvm/pull/6243) | |
| - Implemented REVERSE_V2 Operator for TFLite. [#6304](https://github.com/apache/incubator-tvm/pull/6304) | |
| - Implemented MATRIX_SET_DIAG Operator for Relay/TOPI and TFLite Frontend. [#6303](https://github.com/apache/incubator-tvm/pull/6303) | |
| - RESHAPE with dynamic shape arg in TFLite frontend [#6208](https://github.com/apache/incubator-tvm/pull/6208) | |
| - Constant input attr added to fully connected operation in TFLite frontend [#6228](https://github.com/apache/incubator-tvm/pull/6228) | |
| - Gather operation with indices as tensor expr in TFLite frontend [#6168](https://github.com/apache/incubator-tvm/pull/6168) | |
| - Added support for tflite quantized maximum and minimum [#6018](https://github.com/apache/incubator-tvm/pull/6018) | |
| ### Other frontends | |
| - Unary ops support added in frontend [#6196](https://github.com/apache/incubator-tvm/pull/6196) | |
| - Introduce caffe frontend for tvm [#6206](https://github.com/apache/incubator-tvm/pull/6206) | |
| - Keras softmax and prelu fix under NHWC [#6278](https://github.com/apache/incubator-tvm/pull/6278) | |
| - add support for MXNET numpy operators [#6054](https://github.com/apache/incubator-tvm/pull/6054) | |
| - Refine tensorflow frontend 1.x & 2.x compatibility [#6240](https://github.com/apache/incubator-tvm/pull/6240) | |
| - Reduceops support added to frontend [#6252](https://github.com/apache/incubator-tvm/pull/6252) | |
| - Update precision in the ONNX strided_slice, update precision of ToScalar [#6272](https://github.com/apache/incubator-tvm/pull/6272) | |
| ### TOPI | |
| - Use auto-tuner to improve conv2d_gemm performance [#6117](https://github.com/apache/incubator-tvm/pull/6117) | |
| - topi -> tvm/topi [#6186](https://github.com/apache/incubator-tvm/pull/6186) | |
| ### Build and CI | |
| - TVMC - a command line driver for TVM (Part 1) [#6112](https://github.com/apache/incubator-tvm/pull/6112) | |
| - Remove topi from the CI cache [#6188](https://github.com/apache/incubator-tvm/pull/6188) | |
| - Remove libtopi from the build [#6189](https://github.com/apache/incubator-tvm/pull/6189) | |
| - Update build support for cross compiling apps/cpp_rpc with OpenCL [#6229](https://github.com/apache/incubator-tvm/pull/6229) | |
| - Add docker/lint.sh, for running dockerized lint scripts locally [#6333](https://github.com/apache/incubator-tvm/pull/6333) | |
| - Add gpuonly tests for python unittests and integration [#6346](https://github.com/apache/incubator-tvm/pull/6346) | |
| ### Quantization | |
| - Add Quantize/Dequantize Partitioning [#5940](https://github.com/apache/incubator-tvm/pull/5940) | |
| ### Fixes | |
| - Temporary disable conv2d grad strided flaky test [#6183](https://github.com/apache/incubator-tvm/pull/6183) | |
| - Avoid unexpected throw in AttrInitEntry [#6128](https://github.com/apache/incubator-tvm/pull/6128) | |
| - Fix alignment of note [#6181](https://github.com/apache/incubator-tvm/pull/6181) | |
| - Added casting to hybrid script doc and fixed pass infra doc [#6174](https://github.com/apache/incubator-tvm/pull/6174) | |
| - Fix #6205 [#6207](https://github.com/apache/incubator-tvm/pull/6207) | |
| - Change the meaning of conv3d_transpose output_padding to match conv{1,2}d_transpose [#6065](https://github.com/apache/incubator-tvm/pull/6065) | |
| - Fix compile warnings. [#6204](https://github.com/apache/incubator-tvm/pull/6204) | |
| - Fix -mfloat-abi=soft compilation for ARM with OpenCL target [#6150](https://github.com/apache/incubator-tvm/pull/6150) | |
| - Enable auto conversion String->DLDataType [#6214](https://github.com/apache/incubator-tvm/pull/6214) | |
| - Update pass infra tutorial [#6193](https://github.com/apache/incubator-tvm/pull/6193) | |
| - Mod operator, bug fix [#6160](https://github.com/apache/incubator-tvm/pull/6160) | |
| - Fix compilation error with cuda 11 [#6213](https://github.com/apache/incubator-tvm/pull/6213) | |
| - Fix port_end wrong default value 9199 to 9099 for keeping same with source code [#6220](https://github.com/apache/incubator-tvm/pull/6220) | |
| - Std op without specified dimensions support [#6226](https://github.com/apache/incubator-tvm/pull/6226) | |
| - Fix typo [#6230](https://github.com/apache/incubator-tvm/pull/6230) | |
| - Verify that tensor reshape is valid. [#6215](https://github.com/apache/incubator-tvm/pull/6215) | |
| - Fix crt building and running error [#6231](https://github.com/apache/incubator-tvm/pull/6231) | |
| - Fix conv2d_transpose output padding [#6236](https://github.com/apache/incubator-tvm/pull/6236) | |
| - Fix cuda half math function is undefined: hpow, htanh [#6225](https://github.com/apache/incubator-tvm/pull/6225) | |
| - Split MKL from BLAS. [#6182](https://github.com/apache/incubator-tvm/pull/6182) | |
| - Fix division range estimation error in simplifier [#6244](https://github.com/apache/incubator-tvm/pull/6244) | |
| - Support overriding RPCWatchdog termination behavior on Android and other platforms [#6216](https://github.com/apache/incubator-tvm/pull/6216) | |
| - Revert "fix cuda half math function is undefined: hpow, htanh" [#6249](https://github.com/apache/incubator-tvm/pull/6249) | |
| - Fix newer GCC compiler warnings. [#6257](https://github.com/apache/incubator-tvm/pull/6257) | |
| - Support _contrib_SyncBatchNorm [#6245](https://github.com/apache/incubator-tvm/pull/6245) | |
| - Fix reduction [#6250](https://github.com/apache/incubator-tvm/pull/6250) | |
| - Add apt repository for clang-11 and llvm-11 [#6256](https://github.com/apache/incubator-tvm/pull/6256) | |
| - Update tutorial to new TARGET as micro_dev is no more [#6262](https://github.com/apache/incubator-tvm/pull/6262) | |
| - Improve NHWC depthwise convolution for AArch64 [#6095](https://github.com/apache/incubator-tvm/pull/6095) | |
| - Fix clang-format [#6264](https://github.com/apache/incubator-tvm/pull/6264) | |
| - Trivial fix, up the rodata section for the discovery board to 512 bytes. [#6259](https://github.com/apache/incubator-tvm/pull/6259) | |
| - Fix cuda half math function is undefined: hpow, htanh [#6253](https://github.com/apache/incubator-tvm/pull/6253) | |
| - Add dilation in x86 NCHWc depthwise conv support [#6267](https://github.com/apache/incubator-tvm/pull/6267) | |
| - Decrease test times by introducing testing model [#6235](https://github.com/apache/incubator-tvm/pull/6235) | |
| - Add support for parsing the any dimension. [#6277](https://github.com/apache/incubator-tvm/pull/6277) | |
| - Improve error messages for memory verifier and gpu memory verifier [#6281](https://github.com/apache/incubator-tvm/pull/6281) | |
| - Update ci-cpu to the latest [#6283](https://github.com/apache/incubator-tvm/pull/6283) | |
| - Enable CI for Ethos-N [#6171](https://github.com/apache/incubator-tvm/pull/6171) | |
| - Reflect Compile-Time CMake Options into libtvm.so [#6280](https://github.com/apache/incubator-tvm/pull/6280) | |
| - Add cmake options into libinfo [#6286](https://github.com/apache/incubator-tvm/pull/6286) | |
| - Update slice to infer attributes when not graph inputs [#6276](https://github.com/apache/incubator-tvm/pull/6276) | |
| - Support int4/int8 conv2d tensor core with HWNC layout [#6121](https://github.com/apache/incubator-tvm/pull/6121) | |
| - Use rpc.LocalSession for simple tests [#6294](https://github.com/apache/incubator-tvm/pull/6294) | |
| - Optimize and eliminate the Jacobian tensor for te.autodiff [#6078](https://github.com/apache/incubator-tvm/pull/6078) | |
| - Fix flaky test [#6307](https://github.com/apache/incubator-tvm/pull/6307) | |
| - Multiple output support, reshape, split ops added [#6296](https://github.com/apache/incubator-tvm/pull/6296) | |
| - Fix random fail [#6312](https://github.com/apache/incubator-tvm/pull/6312) | |
| - Fix resize test [#6298](https://github.com/apache/incubator-tvm/pull/6298) | |
| - Fix cython FFI compact with np.int64 [#6321](https://github.com/apache/incubator-tvm/pull/6321) | |
| - Fix relay vm optimize [#6322](https://github.com/apache/incubator-tvm/pull/6322) | |
| - Changed TVMCTVMContext to TVMContext [#6306](https://github.com/apache/incubator-tvm/pull/6306) | |
| - Make able to compile with MSVC [#6341](https://github.com/apache/incubator-tvm/pull/6341) | |
| - ROCm changed name of library and removed the old one in ROCm 3.7 release. [#6345](https://github.com/apache/incubator-tvm/pull/6345) | |
| - Add `init` member to ReduceNode [#6138](https://github.com/apache/incubator-tvm/pull/6138) | |
| - Quanitze operation expanded to take const argument [#6127](https://github.com/apache/incubator-tvm/pull/6127) | |
| - Improve Rust bindings: Map, Array, String, various IR nodes [#6339](https://github.com/apache/incubator-tvm/pull/6339) | |
| - Compatible for ROCm before 3.7 [#6359](https://github.com/apache/incubator-tvm/pull/6359) | |
| - Use clear name that is separate from ASF brand for cache [#6360](https://github.com/apache/incubator-tvm/pull/6360) | |
| - Fix typo [#6352](https://github.com/apache/incubator-tvm/pull/6352) | |
| - Fix mistyped word [#6362](https://github.com/apache/incubator-tvm/pull/6362) | |
| - Fix Dockerfile.demo_android [#6361](https://github.com/apache/incubator-tvm/pull/6361) | |
| - Fix typo [#6338](https://github.com/apache/incubator-tvm/pull/6338) | |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
please add this one to August
@kparzysz-quic -> committer [#6290](https://github.com/apache/incubator-tvm/pull/6290)